55

CPGNNews Strategic Report SR-001/2026 | Author: Yvon Bonhomme | October 2026

CPGNNews — Independent International Public Interest Media | Research • Investigation • Analysis • Journalism.

Born in the Caribbean, Open to the World!

THE AGE OF UNLEASHED INTELLIGENCE

Humanity at a Crossroads of Possibilities: Artificial Intelligence, the Future of Work, and the New Architecture of Global Power

At certain moments in history, humanity develops a technology that does more than transform its tools. It reshapes the way people live, work, produce, communicate, and understand the world around them. The printing press expanded the boundaries of knowledge. The steam engine accelerated industrial transformation. Electricity profoundly altered societies, while computing and the internet brought continents closer together, multiplied exchanges, and redefined relationships among individuals, businesses, and states. Today, another profound rupture is taking shape, one that may prove even more far reaching because it affects not only the capabilities available to humanity, but also certain intellectual capacities, decision making systems, and the very foundations of economic and social organization. This transformation has a name: artificial intelligence.

For the first time on such a scale, systems designed by human beings can analyze vast quantities of information, generate text, images, and audiovisual content, translate languages, assist scientific research, detect certain anomalies, write code, optimize industrial processes, and contribute to decisions once reserved for highly specialized professionals. In laboratories, hospitals, universities, public administrations, businesses, agricultural operations, and households, their applications are multiplying at a pace that often exceeds the capacity of institutions to understand their full implications. What until recently belonged largely to the realm of experimental research is progressively becoming part of everyday life, economic competition, and the strategic calculations of major powers.

Yet the historical significance of this transformation lies not solely in the performance of machines. It also lies in the potential redistribution of capabilities, wealth, and power. Artificial intelligence can enable a small business to access tools once available only to large organizations, help a researcher move more quickly through certain stages of an investigation, allow a teacher to diversify instructional methods, assist a farmer in managing resources more effectively, or enable an isolated community to access services that were previously difficult to obtain. It can also strengthen the position of companies that control the data, processors, computing centers, platforms, and capital required for its development. It can broaden access to knowledge while concentrating the means through which that knowledge is produced and distributed. It can increase individual autonomy while creating new forms of dependency. It can open unprecedented opportunities for less advanced economies, yet also deepen their marginalization when the necessary infrastructure, skills, and investment remain beyond their reach.

This contradiction lies at the heart of one of the defining challenges of our time. Artificial intelligence is neither, by its nature, a universal promise of prosperity nor an inevitable threat to humanity. Its consequences will depend on the scientific, economic, political, and ethical choices that shape its development, on the institutions responsible for governing it, and on societies’ ability to share in the benefits of technological progress. A technology capable of substantially increasing productivity does not guarantee that workers will automatically receive higher incomes. An economy capable of generating greater wealth does not guarantee that this wealth will be distributed more equitably. A society equipped with more powerful information tools does not necessarily become better informed. Nor does a world that is more technologically connected spontaneously become more just, peaceful, or united.

The central question, therefore, is not simply how far machines will advance. It is what kind of future human societies will build with them, who will have the authority to make decisions, who will benefit from productivity gains, who will bear the costs of transition, and who may be left behind. Behind every computational model, industrial investment, automated process, and new application lie choices with direct implications for human interests. These choices concern the dignity of work, freedom of thought, access to education, the protection of personal data, the sovereignty of peoples, institutional accountability, and the place societies wish to reserve for human beings in an increasingly automated environment.

Work is one of the areas in which these questions become most tangible. For millions of people, it represents far more than a source of income. It structures daily life, contributes to individual identity, sustains social relationships, enables the transmission of skills, and gives people the opportunity to participate in collective life. In many families, employment remains the principal safeguard against poverty and one of the few means of attaining a measure of material security. In fragile economies, it can offer a pathway out of precarious circumstances. In ageing societies, it contributes to financing social protection systems. In developing countries, it remains a major determinant of stability, social cohesion, and confidence in the future.

When automated systems begin performing tasks previously carried out by human beings, the transformation extends well beyond questions of efficiency. Certain activities may disappear, others may contract, new occupations may emerge, and existing professions may be reconfigured. Some workers may be freed from repetitive or hazardous tasks, while others may lose part of their autonomy, see their skills devalued, or be compelled to retrain under difficult conditions. The effects will vary across sectors and countries, according to skill levels, public policies, and the ability of businesses to support transitions. It would be just as imprudent to predict the generalized disappearance of work as it would be to promise that everyone will naturally find a place in the new economy. The challenge is to understand the transformations actually taking place, distinguish possibilities from certainties, and prepare societies before the human costs become overwhelming.

This transition also raises questions of intergenerational justice. Children entering school today will probably work, during their professional lives, in occupations whose contours have not yet been fully defined. Students will need to learn how to work with tools capable of producing answers rapidly without abandoning the development of sound judgment, creativity, critical thinking, or the ability to distinguish reliable information from misleading claims. Teachers will need to rethink certain approaches to learning. Policymakers will have to anticipate the implications for employment, taxation, training, and social protection. Businesses will need to determine how to integrate these technologies without treating workers as mere variables to be adjusted. Families, meanwhile, will have to help young people navigate changes affecting education, career prospects, and the ways in which they prepare for their future.

Yet none of these questions can be resolved by looking exclusively at the major industrial economies. The future of artificial intelligence is also being shaped in countries with limited financial resources, incomplete digital infrastructure, or insufficient access to reliable energy. It is taking shape in rural areas where connectivity remains fragile, in schools lacking adequate equipment, in languages with limited digital resources, in small businesses unable to finance their own transformation, and in communities whose needs remain distant from the priorities of major technology markets. It is also taking shape in societies confronted by poverty, natural disasters, institutional crises, and population displacement, where technological promises cannot become reality without fundamental investment in public services and human capabilities.

Africa, the Caribbean, Latin America, parts of Asia, and many other regions nevertheless possess human resources, local knowledge, entrepreneurial capabilities, and cultural traditions that can contribute meaningfully to this transformation. Their future should not be reduced to the role of consumers of technologies designed elsewhere, or to that of suppliers of data, raw materials, or labor. These societies must have the opportunity to participate in research, tool design, standard setting, and value creation. They must be able to develop applications adapted to their languages, economies, and social priorities. They must also have the means to protect their data, educate and train their populations, and negotiate technological partnerships without surrendering their fundamental interests. The digital divide is no longer limited to the difference between those who have an internet connection and those who do not. It increasingly concerns the ability to produce technology, understand it, adapt it, and participate in the decisions that determine how it is used.

The geopolitical dimension of this revolution is equally decisive. States that control computing infrastructure, semiconductor supply chains, capital, telecommunications networks, and the most advanced technologies acquire significant economic and strategic advantages. Rivalries among major powers are no longer confined to territories, natural resources, or traditional military capabilities. They now extend to control over data, technical standards, cybersecurity, satellite systems, energy, and the terms of access to digital infrastructure. In this environment, technological dependence can become an economic vulnerability, an instrument of diplomatic pressure, or a source of fragility during periods of crisis. Sovereignty does not necessarily mean producing everything domestically. It does, however, require an understanding of existing dependencies, diversification of partners, protection of essential interests, and the preservation of genuine decision making capacity.

This competition is not unfolding in an immaterial space. Artificial intelligence depends on equipment, mines, factories, undersea cables, data centers, electricity grids, water, land, and global logistics networks. The infrastructure that makes digital services possible carries an energy and environmental footprint that would be irresponsible to ignore. As computing requirements increase, societies will have to weigh the expansion of technological capabilities against electricity availability, water resource management, emissions reduction, and the other essential needs of their populations. An innovation cannot be evaluated solely in terms of its computing power or financial profitability. It must also be assessed in light of its material costs, social consequences, and actual contribution to human well being.

These challenges are compounded by questions of truth, trust, and accountability. Artificial intelligence systems can facilitate access to information and help analyze complex subjects. They can also generate convincing errors, reproduce biases, amplify misleading content, and make it more difficult to establish the origin of a document, image, or recording. In an environment where information influences elections, markets, international relations, and perceptions of conflict, the ability to verify facts becomes an essential condition of collective freedom. The power of these tools does not relieve those who design, market, use, or disseminate their outputs of responsibility. On the contrary, it makes a culture of verification, transparency, and discernment indispensable.

Nevertheless, it would be reductive to view this revolution solely as a succession of risks. Artificial intelligence can help accelerate certain scientific discoveries, improve disease detection, support research into new materials, optimize energy consumption, facilitate learning, and strengthen preparedness for certain disasters. It can help public administrations organize services more effectively, enable entrepreneurs to test ideas at lower cost, and offer people with disabilities new ways to access information and communicate. These possibilities are neither automatic nor equally accessible to everyone. They nevertheless demonstrate that the choice is not between progress and its rejection, but rather between different ways of ensuring that progress becomes useful, responsible, and broadly shared.

This is precisely why artificial intelligence must be understood as a question of civilization. It challenges the meaning societies assign to development, the place they reserve for knowledge, their conception of justice, and their ability to organize prosperity that is not measured exclusively by the accumulation of wealth. If productivity gains make it possible to produce more while requiring fewer human hours for certain tasks, societies will have to decide how those gains should be distributed, how training should be financed, how people affected by transitions should be protected, and how contributions that cannot be reduced to economic output should be recognized. If machines can perform an increasing share of intellectual work, human worth cannot be reduced to the tasks a person can perform more quickly or cheaply than an automated system. Human dignity does not depend on competing with machines.

Why CPGNNews Chose to Examine This Transformation

CPGNNews has chosen to devote an in depth study to artificial intelligence because this transformation stands at the intersection of nearly every major issue shaping the future of human societies. It simultaneously affects the economy, employment, education, science, culture, the media, governance, security, international relations, and state sovereignty. It changes the conditions under which wealth is generated, knowledge is acquired, decisions are made, and power relationships are established. No international media organization seeking to understand the defining transformations of its time can therefore treat artificial intelligence as a peripheral subject or as a mere succession of technological innovations. It must examine it as a force for change whose consequences extend far beyond the digital sector and affect humanity as a whole, albeit to varying degrees.

For CPGNNews, this editorial choice also reflects a specific responsibility: to help the public understand what lies behind spectacular announcements, massive investments, and commercial promises. The performance of a new model, the industrial projects of major companies, and the technological strategies of global powers acquire their full significance only when examined within their economic, social, human, and geopolitical context. Who controls the infrastructure essential to this revolution? Who possesses the capacity to conduct research and manufacture advanced technologies? Who can finance the adaptation of an economy? Which occupations will be transformed, which skills will become essential, and which populations risk being left behind? How will productivity gains be distributed? What safeguards will protect people’s rights, dignity, and autonomy? These questions, far more than fascination with technology itself, justify sustained and rigorous journalistic investigation.

This approach is particularly important for a media organization born in Haiti, rooted in the Caribbean, and open to the world. The region cannot afford to discover too late the consequences of a transformation that will influence the skills in demand, investment flows, international trade, education systems, and development prospects over the coming decades. Countries with substantial financial resources and advanced infrastructure are already seeking to strengthen their technological capabilities. Societies with more limited resources must also be able to understand the choices available to them, identify the risks of dependence, and prepare their populations for the changes ahead. For Haiti, as for other Caribbean societies, the challenge is not to reproduce the models of major powers mechanically, but to determine how new technological capabilities could contribute to education, vocational training, entrepreneurship, agriculture, healthcare, institutional resilience, and the creation of economic opportunities adapted to local realities.

The Caribbean perspective is therefore not a geographical limitation on this study. On the contrary, it strengthens its universal relevance. Examining artificial intelligence from a region exposed to development inequalities, economic vulnerabilities, energy constraints, and limited access to infrastructure makes it possible to ask a question that discussions centered on major powers sometimes overlook: under what conditions can this revolution benefit those who do not yet possess the means to direct it? An innovation becomes genuine shared progress only when its possibilities can be understood, adapted, and used beyond the centers where capital and decision making capabilities are concentrated. The world cannot sustainably build its technological future while leaving part of humanity in a position of permanent dependence.

By choosing this subject, CPGNNews seeks to fulfill its mission: to inform beyond immediate news, analyze beyond announcements, and connect decisions made in major centers of power with the realities experienced by populations. The objective is neither to celebrate artificial intelligence unreservedly nor to fuel disproportionate fears. It is to subject its promises to factual scrutiny, assess its risks rigorously, distinguish projections from demonstrated results, and make the choices that will determine its use more understandable to the public. Editorial independence demands precisely this critical distance: not confusing innovation with progress, power with justice, growth with well being, or technical capability with human responsibility.

This study also rests on a fundamental conviction: major technological transformations must be assessed not only in terms of what machines can accomplish, but also in terms of what societies choose to preserve, develop, and share. CPGNNews believes that the debate over artificial intelligence concerns every people, every generation, and every individual, even when its mechanisms appear distant or difficult to understand. It concerns the worker wondering about the future of a profession, the student preparing for a career in transition, the teacher adapting instructional methods, the entrepreneur seeking to innovate, the public official required to make informed decisions, and the citizen seeking to understand the forces shaping the future.

This is why CPGNNews has chosen to devote this report to the age of artificial intelligence: not to follow a trend, but to illuminate a transformation that will shape humanity’s economic, social, and political future. Born in the Caribbean, the media organization embraces its international vocation, guided by a simple conviction: understanding the world taking shape is essential if peoples are to claim their place in it, defend their dignity, and participate in the decisions that will shape their future.

A Collective Responsibility in the Face of a New Form of Power

The challenge is twofold. First, we must rigorously understand the technological, economic, and geopolitical transformations under way, without yielding to sensational predictions or minimizing changes that are already observable. At the same time, we must confront the political and moral question running through them all: in whose interests will this new power be organized? A technology with such far reaching implications cannot be left solely to market decisions, any more than it can be governed effectively through fear, improvisation, or prohibitions disconnected from reality. It calls for cooperation among researchers, businesses, workers, educators, governments, media organizations, civil society groups, and citizens. It requires institutions capable of acting at multiple levels, from local communities to international forums, while respecting fundamental rights, the diversity of societies, and the need for clearly defined accountability.

This report proceeds from the conviction that the future of artificial intelligence will be determined not only by those who build the most powerful machines, but also by societies’ ability to define human centered objectives, share knowledge, and prevent the benefits of this revolution from becoming the exclusive privilege of a minority. It examines the emerging balance of power surrounding infrastructure and major corporations, the potential transformation of work, the promise of productivity gains, the risks of economic concentration, digital divides, sovereignty choices, and the conditions required for less advantaged countries to participate in this new chapter of history. Its purpose is neither to celebrate every innovation unconditionally nor to proclaim a predetermined future of decline. It is to understand what is changing, what remains uncertain, and what today’s decisions may mean for generations to come.

For behind the billions invested, the data centers, processors, algorithms, and rivalries among global powers lies a reality that the language of technology can sometimes obscure: human beings hoping to live with dignity, learn, work, protect their loved ones, participate in society, and pass on a better future to their children. The ultimate question is therefore not whether artificial intelligence will become more powerful, but whether humanity will become more responsible as that power grows. Progress fully deserves its name only when it expands human possibilities without sacrificing dignity, creates wealth without abandoning the most vulnerable, and brings peoples closer together rather than turning their differences into new lines of exclusion.

This is the defining challenge of our century: to ensure that artificial intelligence becomes not merely a revolution in power, but an advance in civilization; not wealth created by a few for a few, but a new capability placed at the service of all humanity. Technology can redraw the world, but it is for humanity to decide what kind of world it wishes to inhabit.

THE MASTERS OF ARTIFICIAL INTELLIGENCE

1. The New Global Power No Longer Resides Solely in Territory

For centuries, the history of great powers was told in terms of territorial possession, control of trade routes, access to natural resources, and military superiority. The Industrial Revolution added mastery of machinery, energy, and production capacity to this equation. The twentieth century subsequently established the importance of scientific research, electronics, aerospace, telecommunications, and nuclear weapons. In the twenty first century, another transformation is underway: power increasingly depends on the ability to design, train, deploy, and control artificial intelligence systems capable of influencing economic production, the flow of information, scientific research, administrative decision making, and military operations. This evolution does not make territory, natural resources, or armed forces obsolete. Rather, it changes the conditions under which these traditional instruments of power can be mobilized.

Artificial intelligence is not an isolated sector that can be assessed solely by evaluating the quality of a software product or the performance of a conversational model. It constitutes an industrial and strategic system that brings together semiconductors, data centers, computing infrastructure, cloud services, telecommunications networks, electricity, cooling systems, data, capital, and human expertise. This physical architecture is complemented by mathematical models, algorithms, specialized software, governance frameworks, and commercial arrangements that determine who can access available capabilities and under what conditions. A country may have outstanding researchers without possessing the industrial capacity needed to manufacture the most advanced chips. A company may develop a high performing model without controlling the data centers that host it. A government may announce an ambitious strategy without having the electrical grid, financial resources, or professional expertise required to implement it.

It is in the gap between theoretical capability and effective control of infrastructure that a substantial part of the new global hierarchy is taking shape. Power no longer consists solely of inventing a technology. It also requires control over enough elements of its value chain to deploy it, improve it, finance it, and make it accessible without depending entirely on decisions made elsewhere. The issue also concerns continuity of service. A company or state that depends on a single provider may discover, when a crisis occurs, that its autonomy was more contractual than real. Artificial intelligence is therefore exposing economic and strategic dependencies that already existed, while creating new ones of its own.

2. The United States and the Architecture of a Dominant Ecosystem

The United States occupies a central position in the global artificial intelligence economy because it brings together several mutually reinforcing advantages. The country is home to major technology companies, financial markets capable of supporting enormous investments, leading universities and research laboratories, a substantial domestic market, and an ability to attract researchers and entrepreneurs from many countries. This concentration of resources facilitates the creation of an ecosystem in which scientific innovations can rapidly become commercial products and subsequently be integrated into services used around the world. American dominance therefore does not rest on a single actor, but on the interconnected relationships among model developers, hardware suppliers, data center operators, software developers, cybersecurity companies, and digital distribution networks.

Companies such as Microsoft, OpenAI, Google, Amazon, and Meta occupy different but complementary positions within this architecture. Microsoft combines its Azure cloud services and professional tools with its artificial intelligence partnerships. OpenAI develops models and products for both the general public and organizations. Google brings together research capabilities, proprietary computing infrastructure, digital services, and its Gemini models. Amazon combines a major cloud hosting business with investments in chips and artificial intelligence systems. Meta invests in its own models and in open approaches that can facilitate the adoption of certain technologies by developers. Apple follows a different strategy, focusing more heavily on integrating artificial intelligence into its devices, software, and data processing systems. These strategies are not identical, and the companies compete vigorously with one another, but they all illustrate the same underlying reality: value is created through the combination of software, hardware, infrastructure, data, distribution, and relationships with users.

The cooperation between Microsoft and OpenAI also illustrates how power relations within this ecosystem are evolving. A strategic partnership can provide a research laboratory with computing capacity, financing, and commercial distribution while creating significant dependence on its industrial partner. Developments announced in April 2026 concerning their relationship demonstrate that these arrangements are not immutable and that the parties seek to preserve room for maneuver as their activities expand. The issue extends beyond the terms of a contract. It concerns access to infrastructure, intellectual property, product distribution, and the ability to work with multiple providers. In an industry characterized by exceptionally high investment requirements, technological autonomy is often negotiated through complex partnerships rather than achieved through absolute independence.

The concentration of capabilities nevertheless has its limits. The expenditure required to train the most advanced models, access to specialized chips, and the availability of electricity can reinforce the advantages of companies that are already well established. Smaller businesses, universities, and countries with limited resources may consequently become dependent on services they cannot reproduce or fully control. Open models can lower certain barriers to access and allow developers to modify existing systems, but they do not automatically eliminate the need for computing power, expertise, financing, and infrastructure. It is therefore essential to distinguish between an open model and the autonomy of the ecosystem that makes it usable.

3. Semiconductors: The Material Foundation of Artificial Intelligence

Behind the interfaces available to the public lies a far more complex industry: semiconductors. Specialized artificial intelligence processors perform the mathematical operations required to train and run models at high speed. Their design, manufacturing, and integration require extremely expensive industrial equipment, high purity materials, specialized scientific expertise, and exceptionally precise supply chains. The strategic value of chips derives not only from their raw processing power, but also from their energy efficiency, their ability to communicate with one another, their memory capacity, and their suitability for operating within large scale systems.

NVIDIA has become a central player in this environment through its processors and software ecosystem, which facilitate the development and execution of numerous artificial intelligence applications. Other companies are developing their own chips to reduce costs, improve performance, or limit their dependence on particular suppliers. Amazon, for example, invests in processors designed for its own infrastructure. This diversification does not mean that industrial barriers are disappearing. Designing a powerful chip does not guarantee that it can be manufactured at scale, at a competitive cost, and with satisfactory production yields. The industry depends on an internationally specialized structure in which design expertise, lithography equipment, advanced manufacturing, and assembly are distributed among multiple players.

Taiwan occupies a pivotal position in this structure, particularly through Taiwan Semiconductor Manufacturing Company, known as TSMC, which manufactures advanced chips for numerous international designers. This position makes the island an essential industrial hub for the global digital economy. It also leaves the supply chain vulnerable to any major crisis in the Taiwan Strait. A prolonged disruption could affect electronics manufacturers, data centers, telecommunications, the automotive industry, and several defense sectors. Inventories, production capacity located elsewhere, and diversification strategies can mitigate some risks, but they may not make it possible to replace all the most advanced manufacturing capabilities quickly.

ASML, the Dutch company specializing in advanced lithography systems, occupies a strategic position in the production of the most sophisticated chips. Its role demonstrates that technological sovereignty cannot be reduced to owning factories or examining the nationality of companies. It depends on a combination of expertise, equipment, and suppliers, none of whom necessarily controls the entire process. The United States, Europe, Taiwan, South Korea, Japan, and other economies participate in different parts of this chain. Export restrictions, industrial policies, and security agreements have consequently become instruments of technological competition. They can protect certain strategic advantages, but they can also generate economic costs, accelerate substitution strategies, and fragment markets.

4. Terafab: An Attempt to Integrate More Links in the Industrial Chain

The Terafab project announced in August 2026 by Elon Musk, in association with Tesla and SpaceX, illustrates an ambition that is particularly revealing of this new competition. Presented as an industrial complex in Grimes County, Texas, the project aims to bring together several operations that are generally distributed among different companies, including logic and memory chip manufacturing, assembly, packaging, and testing. The envisaged applications include embedded inference systems, Optimus robots, autonomous vehicles, and certain computing infrastructures associated with space activities. SpaceX has stated that the combined demand of Tesla and SpaceX for chips and computing capacity could exceed one terawatt of computing. This figure describes the demand anticipated by the two companies; it does not represent an announced annual production capacity for Terafab. It is a projection made by the project’s promoters, not an already operational capability or a demonstrated level of industrial output.

The project addresses a strategic concern: companies seeking to deploy robots, autonomous vehicles, and computing infrastructure on a massive scale may regard access to chips as a limiting factor. By investing in their own industrial capacity, they would seek to coordinate hardware design, software requirements, production volumes, and deployment schedules more effectively. Successful vertical integration could give them greater control over costs and component availability. It could also reduce certain dependencies on third party manufacturers. However, such a strategy requires the resolution of substantial industrial challenges, including mastery of manufacturing processes, access to specialized equipment, recruitment of scarce expertise, production line yields, component quality, and continuous investment in successive generations of technology.

Details reported in October 2026 indicate that Tesla and SpaceX intend to build and operate the project. Elon Musk has also clarified that TSMC could lease part of the facilities without becoming their operator. This distinction matters because an industrial announcement, even when accompanied by an initial investment and a proposed timeline, does not establish that its promoters will master the most advanced manufacturing technologies at the intended scale. Announcing an integrated factory is not sufficient to demonstrate that its processes can immediately operate with the yields, costs, and quality standards required. Industrial experience, equipment, manufacturing processes, and production yields are developed over time. Terafab should therefore be analyzed as an ambitious attempt to change the organization of semiconductor production, not as proof that existing constraints have already been overcome.

The project also highlights a broader evolution in technological capitalism. Major companies no longer want merely to purchase components in a specialized market. They sometimes seek to secure their own industrial capabilities when they consider them essential to their strategies. This trend can encourage innovation and improve coordination between hardware and software, but it also entails financial and operational risks. The deeper the integration, the more responsibility a company must assume for activities previously handled by specialized suppliers. Terafab will therefore be judged less by the scale of its announcements than by its ability to produce reliable, competitive components in sufficient quantities.

THE MASTERS OF ARTIFICIAL INTELLIGENCE

5. The Rivalry Between Washington and Beijing and the Role of Taiwan

Competition between the United States and China is one of the principal geopolitical dimensions of artificial intelligence. It encompasses models, processors, computing infrastructure, telecommunications, industrial applications, technical standards, and the ability to attract talent. Washington seeks to preserve its advantages in several critical technologies, while Beijing is investing in domestic capabilities to reduce its dependence on foreign suppliers. China has a vast domestic market, a substantial industrial base, significant scientific and technical human resources, and major digital companies. However, it faces particular constraints in certain segments of the advanced semiconductor industry, partly because of trade restrictions and the difficulty of rapidly reproducing all the most sophisticated equipment and manufacturing processes.

It would nevertheless be misleading to reduce China’s strategy to a simple attempt at imitation. Chinese companies and institutions are developing their own models, infrastructure, and applications adapted to their economic environment. American restrictions may slow certain activities, but they also encourage investment in domestic solutions and alternative supply chains. The outcome will depend on China’s ability to improve performance, costs, industrial yields, and access to energy, as well as on the effectiveness of its research and training policies. Likewise, the United States cannot regard its lead as permanently secure. Competition is evolving rapidly, and technological advantages can narrow when a rival discovers more efficient methods or adopts a different architecture.

Taiwan lies at the heart of this equation because of its industrial importance and political status. The island has its own institutions and system of government, while the People’s Republic of China claims it and has not renounced the possibility of using force. The United States maintains unofficial relations with Taiwan and provides substantial support, particularly in defense, within the framework of its legislation and policy toward the Taiwan Strait. This configuration creates a complex deterrence environment in which military decisions, diplomatic signals, alliances, and economic calculations interact. A crisis would not be merely a regional confrontation. It could trigger a global shock affecting electronics production, financial markets, trade, and digital infrastructure.

Taiwan’s centrality thus creates a strategic contradiction. Economies that depend on its industrial capabilities have an interest in preserving stability in the Taiwan Strait, yet security policies and sovereignty disputes can intensify tensions. Geographical diversification of manufacturing capacity can reduce certain vulnerabilities without rapidly eliminating the importance of Taiwan’s ecosystem. The challenge is therefore to build more resilient supply chains while preventing technological competition from turning a shared economic dependency into an instrument of political escalation.

6. Satellite Networks and the Transformation of Warfare

Technological power is not confined to terrestrial data centers. It also depends on the ability to transmit information, maintain communications in degraded environments, and connect remote regions. Low Earth orbit satellite constellations have transformed access to the Internet, particularly in areas poorly served by traditional networks. They can complement terrestrial infrastructure, accelerate the restoration of communications after a disaster, and provide useful connectivity in rural or isolated areas. Their importance has become especially visible in armed conflicts, where communication networks are essential infrastructure for coordinating operations and maintaining civilian services.

In Ukraine, Starlink, SpaceX’s satellite service, has played an important role in maintaining certain communications following Russia’s full scale invasion. This experience has demonstrated that privately owned digital infrastructure can, under certain circumstances, acquire strategic importance comparable to that of equipment traditionally controlled by states. It has also exposed a major institutional question: when a private capability becomes indispensable to military operations or to the continuity of a country’s communications, the commercial and technical decisions of its operator can have consequences extending far beyond its ordinary customers. The issue is therefore not limited to network performance. It concerns access conditions, usage restrictions, contractual responsibilities, coordination mechanisms with governments, and alternative solutions in the event of an interruption.

Russia is seeking to develop alternative satellite capabilities. Information published in June 2026, in particular, described a Russian project intended to launch a satellite broadband service in 2027 on a more limited scale than Starlink. This announcement indicates a strategic intention, but it does not demonstrate that an equivalent network is already operational or that it can provide comparable coverage, capacity, and resilience. Building a constellation requires satellites, regular launches, user terminals, ground stations, financial resources, and an industrial supply chain capable of sustaining the system over time. A network’s effectiveness cannot be measured solely by the number of satellites announced. Its availability, throughput, latency, coverage, and resistance to disruption are equally important.

European concerns reported in October 2026 regarding Russian satellite activities near commercial systems used by Western partners also illustrate the growing importance of space surveillance. These reports must be attributed to the officials who made them and must not be presented as proof of a confirmed attack without additional evidence. Nevertheless, they reflect growing concerns about commercial satellites, intelligence gathering, and the security of orbital infrastructure. Space is gradually becoming an arena in which military interests, civilian communications, Earth observation, and economic activities intersect.

The lesson is clear: digital sovereignty depends on more than the ability to store data within national borders or own a terrestrial network. It also requires an understanding of dependencies on orbital infrastructure, foreign suppliers, and private operators. For small states and island territories, satellites can offer considerable opportunities, particularly for connectivity and resilience after disasters. Their use, however, should form part of a policy of diversification and service continuity rather than create exclusive dependence on infrastructure whose decisions lie entirely beyond the control of local authorities.

THE MASTERS OF ARTIFICIAL INTELLIGENCE

7. Europe and the Challenge of Technological Sovereignty

The European Union possesses substantial scientific, industrial, and financial resources. It is home to internationally recognized universities, high technology companies, expertise in industrial equipment, and institutions capable of supporting large scale research policies. ASML’s role in the production of advanced semiconductor manufacturing equipment demonstrates that Europe can occupy a critical position in the global technology supply chain. Companies such as Mistral AI are also contributing to efforts to develop European models and capabilities. The problem, therefore, is not a lack of expertise or initiatives, but the difficulty of transforming these assets into an ecosystem that is sufficiently integrated, well funded, and widely deployed to compete globally with major American platforms and China’s industrial capabilities.

European technological sovereignty faces several constraints. Markets remain fragmented by languages, regulatory frameworks, national policies, and differences in access to financing. Innovative companies may struggle to move from research to industrial deployment and then to global commercialization. Europe must also reconcile its digital ambitions with energy costs that weigh on certain industrial activities, substantial investment requirements, and difficult trade offs between competitiveness, social protection, and the climate transition. Debates over immigration, identity, unemployment, public services, and the distribution of transition costs can make it harder to build lasting political consensus. Yet it would be misleading to interpret these tensions as evidence of complete paralysis. European institutions and member states continue to develop industrial policies, partnerships, and economic security initiatives, with varying results across different fields.

The military dimension adds another layer of complexity. European countries have armed forces, defense industries, and national capabilities that should not be overlooked. Nevertheless, they remain dependent on the United States for certain essential capabilities within the transatlantic security architecture, particularly in areas such as intelligence, surveillance, strategic transport, some command functions, and several defense systems. This dependence does not mean that Europe lacks military resources. It means that Europe might not be able to replace all American capabilities immediately if political or military circumstances were to change fundamentally. The declaration adopted at the NATO summit held in Ankara in July 2026 underscores the growing importance of burden sharing among European allies, Canada, and the United States, while maintaining the framework of the transatlantic alliance.

This situation has direct implications for artificial intelligence. Technologies used in defense, intelligence, cybersecurity, and communications increasingly rely on many of the same industrial infrastructures as civilian applications. The ability to protect data, maintain operational systems, and develop solutions tailored to military requirements is becoming an essential component of sovereignty. Europe must therefore reconcile several objectives: maintaining its partnerships, developing its own capabilities, avoiding inefficient duplication, and preserving access to technologies for which it does not yet control the entire production chain. Sovereignty does not mean autarky. It means having credible alternatives, reducing critical dependencies, and retaining the ability to negotiate without being entirely exposed to the decisions of a single partner.

8. The Middle East: Turning Financial Resources into Technological Power

Several Gulf countries, particularly the United Arab Emirates and Saudi Arabia, are investing in digital infrastructure, data centers, artificial intelligence, and international technology partnerships. Their financial resources, ability to mobilize capital rapidly, and access to energy resources allow them to contemplate projects on a considerable scale. These investments form part of broader economic diversification strategies intended to reduce dependence on hydrocarbons and develop higher value added industries. They also reflect a desire to use digital technology as an instrument of diplomatic influence, administrative modernization, and industrial development.

However, access to capital and energy does not guarantee complete technological autonomy. A country may finance a data center without mastering the manufacture of its processors, the software that powers its systems, or the skills required to maintain and upgrade them. It may conclude agreements with major companies without possessing all the leverage needed to negotiate access conditions, data localization, permitted uses, or service continuity. The challenge is to turn investment into national expertise, research capacity, maintenance capabilities, local enterprises, and lasting economic value. Success will depend as much on the quality of governance and education as on the volume of capital mobilized.

The Middle East must also not be treated as a homogeneous region. Resources, institutions, economic priorities, political circumstances, and scientific capabilities differ considerably from one country to another. Artificial intelligence strategies can improve certain public services and support industrial sectors, but they also raise questions about surveillance, data protection, employment, and the distribution of benefits. The Gulf experience therefore offers lessons for other economies seeking to accelerate their digital transformation: investment is necessary, but its strategic value depends on whether it is embedded in a broader policy of human development and institutional control.

9. Developing Countries and Small States: Access Is Not the Same as Sovereignty

Developing countries are approaching the artificial intelligence revolution from very different starting points. Some have universities, technology companies, and expanding digital markets. Others continue to face substantial deficits in electricity, connectivity, financing, and training. Even when applications are available free of charge or at low cost, productive use requires suitable devices, reliable connectivity, digital skills, and procedures for verifying results. A public administration may use an artificial intelligence tool to improve drafting or document processing without having a data governance strategy, cybersecurity safeguards, or mechanisms for overseeing automated decisions. Access to a tool is therefore a first step, but it must not be confused with the ability to choose, adapt, control, and maintain it.

Yet the needs of developing countries are considerable. In agriculture, image analysis and weather data can help detect certain crop diseases, improve planting schedules, and make better use of water. In education, assistance systems can complement teachers’ work, support language learning, and provide exercises adapted to different skill levels. In healthcare, decision support tools can contribute to triage and information management, provided that medical limitations are respected and professional supervision is maintained. Public administrations can also use these technologies to improve access to information, translate documents, identify inconsistencies, and accelerate repetitive tasks. These opportunities cannot replace institutions, professionals, or basic investment. They can, however, strengthen existing capabilities when deployment responds to clearly identified needs.

Small island developing States face additional constraints. Their geographic isolation, exposure to natural disasters, small markets, and dependence on essential imports can increase the cost of digital infrastructure. Hurricanes, floods, earthquakes, and network disruptions can undermine service continuity precisely when communication needs are most urgent. Artificial intelligence can contribute to risk mapping, weather data analysis, emergency preparedness, and the dissemination of useful information. Its effectiveness, however, depends on data quality, equipment maintenance, and proven institutional protocols. Advanced technology cannot compensate for fragile infrastructure; it must be integrated into a broader resilience strategy.

The digital sovereignty of these countries should not be understood as an obligation to manufacture every technological component themselves. Such a requirement would be financially unrealistic for many of them and could lead to inefficient duplication. Sovereignty instead requires the ability to make informed choices: compare providers, avoid excessive dependence on a single actor, protect sensitive data, train users, negotiate transparent contracts, and maintain continuity arrangements. Regional cooperation can play a decisive role by pooling training programs, research infrastructure, cybersecurity capabilities, language resources, and tools adapted to local needs. For small economies, shared capabilities may provide greater practical autonomy than isolated attempts to reproduce the entire global industrial supply chain.

10. Haiti and the Caribbean: Making Technology an Instrument of Development

Haiti and the Caribbean countries cannot afford to approach artificial intelligence solely as an external phenomenon whose consequences they simply observe. The region faces structural difficulties that limit its ability to translate innovation rapidly into productivity gains: unequal access to electricity, connectivity that can be expensive or unreliable, weaknesses in certain public services, substantial training needs, vulnerability to disasters, and limited access to long term capital. These constraints are compounded by a linguistic challenge: digital resources do not always adequately reflect the region’s realities, languages, and social contexts. These obstacles are serious, but they do not make innovation impossible. Rather, they show why technology choices must be guided by concrete priorities and a realistic assessment of implementation conditions.

In agriculture, analytical tools can help monitor crops, manage water resources, distribute practical advice, and improve planning. In education, properly supervised applications can support teachers, strengthen learning in Haitian Creole, French, English, and Spanish, and improve access to educational resources. In disaster risk management, analysis of available data can help map vulnerable areas, organize disaster related information, and improve emergency communications. In public administration, certain tools can simplify document research, translation, file classification, and report preparation. None of these applications, however, can replace investment in schools, electricity networks, telecommunications, vocational training, and the institutions responsible for the quality of public services.

Language is a particularly important issue. When tools are primarily designed and evaluated in languages supported by extensive digital corpora, speakers of less represented languages may receive results of uneven quality. Haitian Creole must not be treated as a mere translation issue. It is central to access to information, education, civic participation, and the ability of institutions to communicate with the population. Developing reliable language resources, high quality corpora, and tools adapted to local circumstances can help reduce certain forms of digital exclusion. This ambition nevertheless requires respect for the rights of individuals whose data are used, careful verification of generated content, and a clear distinction between a fluent answer and a reliable one.

The Caribbean can also benefit from regional cooperation and its diaspora networks. Universities, professional organizations, media outlets, businesses, and public institutions could share training programs, language resources, disaster management tools, and methods for verifying information. Countries in the region do not need to reproduce the most expensive infrastructure independently in order to develop collective capabilities for using and governing these technologies. They should instead identify areas in which joint action would produce tangible results and strengthen their negotiating position with international providers.

In this environment, CPGNNews can contribute to public understanding of technological change by examining industrial decisions, the strategies of major powers, social consequences, and regional development opportunities. Its mission as an international public interest media organization, born in the Caribbean and open to the world, means that its coverage must not be confined to traditional centers of power. The Caribbean is the point of origin, not the limit. This perspective makes it possible to examine how global transformations affect island societies while giving those societies a place in debates about the technological and economic choices that will shape the future.

11. Energy, Water, and the Material Limits of Digital Expansion

The expansion of artificial intelligence depends on physical infrastructure whose costs are becoming increasingly visible. Data centers consume electricity to power servers, storage systems, and network equipment. They must also remove the heat generated by processors, requiring cooling systems whose needs vary according to facility design, climate, and the technologies used. Building additional computing capacity can therefore place pressure on electricity grids, water resources, and industrial equipment. The scale of this pressure depends heavily on the applications involved, infrastructure efficiency, energy sources, and the rate at which demand grows.

It would nevertheless be inaccurate to conclude that every use of artificial intelligence automatically consumes more energy than traditional methods. Some applications can reduce travel, improve industrial planning, or optimize operations that previously required greater resources. Assessment must cover the entire life cycle, including hardware manufacturing, operation, service life, and replacement. The rebound effect must also be considered: when a technology becomes cheaper or easier to use, demand may rise enough to offset part of the expected savings. Technical efficiency remains essential, but it does not by itself guarantee a reduction in total consumption.

Materials and industrial supply chains raise further concerns. Electronic equipment requires metals, chemicals, and demanding manufacturing processes. Rapid replacement cycles can increase electronic waste and make recycling more difficult. Governments and businesses must therefore treat equipment durability, repairability, waste management, and diversified sourcing as components of their digital strategies. Countries that import most of their equipment are particularly vulnerable to price fluctuations, logistical disruptions, and supplier decisions.

For Haiti and other vulnerable economies, these constraints make it essential to prioritize solutions proportionate to actual needs. Strengthening electricity reliability, connectivity in schools and training centers, or the security of administrative systems may be more useful than immediately launching large computing projects whose operating costs would be difficult to sustain. Digital investment should be assessed according to its social benefits, financial sustainability, and compatibility with available infrastructure. Modernization should not be measured by the size of the equipment announced, but by the ability to deliver lasting benefits to the population.

12. Who Will Control the Value Created by Artificial Intelligence?

Technological competition concerns not only the creation of value but also its distribution. If artificial intelligence enables a company to produce more with fewer resources, the gains may take the form of higher profits, lower prices, new products, improved wages, or workforce reductions. There is no guarantee that these effects will be distributed automatically or equitably. Companies that control infrastructure, models, data, and distribution channels may have greater bargaining power than workers, small businesses, and local suppliers. The central question is therefore political and economic: who benefits from productivity gains, who bears the costs of adjustment, and what mechanisms can make the transformation acceptable and sustainable?

The answers depend on markets, institutions, and public policy. Effective competition can limit certain forms of economic rent, while interoperability and open standards can make it easier to switch providers and reduce the costs of dependency. Data protection rules can limit abusive practices, provided that they are enforceable and proportionate. Tax policy, training programs, social dialogue, and worker protection measures can help distribute transition costs. In countries with limited administrative capacity, implementing such policies may be difficult. This makes clear project governance and cooperation among public institutions, businesses, and civil society organizations all the more important.

The issue also affects the media and the circulation of information. Generative systems can help locate documents, compare data, translate content, and automate certain production tasks. They can also produce plausible but incorrect information, reproduce biases in their training data, or facilitate the spread of misleading content. Organizations using these systems must therefore maintain verification procedures, protect their sources, and clearly assign editorial responsibility. In public interest journalism, these tools can strengthen investigative work, but they cannot replace editorial judgment, independent verification, or human accountability to readers.

The real question is not whether artificial intelligence will be used, since its adoption is already underway across many sectors. The question is under what conditions it will be used, what safeguards will apply, and who will benefit. Technology choices affect employment, competition, sovereignty, civil liberties, and development opportunities. A society that simply adopts tools without defining its priorities risks deepening its dependencies. A society that invests in skills, transparency, infrastructure, and bargaining capacity can instead turn this revolution into an instrument of progress.

THE TRUTH BEHIND THE NUMBERS

13. Measuring Exposure Does Not Mean Predicting the Disappearance of Jobs

Debates about artificial intelligence and work are often dominated by dramatic figures. Hundreds of millions of jobs are said to be exposed to automation, millions of workers may need to change occupations, and substantial productivity gains could transform the global economy. These estimates matter, but they do not all measure the same thing. Some assess the proportion of tasks that can be performed or assisted by an artificial intelligence system; others model the possible elimination of certain activities, occupational transformation, or the creation of new jobs. Their reference periods, populations, technological assumptions, and calculation methods also differ. Comparing these figures without examining their definitions risks confusing economic scenarios with certain predictions.

A report published in May 2025 by the International Labour Organization and NASK, Poland’s national research institute, estimates that approximately one in four jobs worldwide has some degree of exposure to generative artificial intelligence. The proportion rises to about 34% in high income countries, compared with 11% in low income countries. The study also estimates that the global category with the highest degree of exposure represents approximately 3.3% of total employment. Women are more represented than men in this category, accounting for 4.7% compared with 2.4% globally. In high income economies, the gap is more pronounced in this category, with 9.6% of women compared with 3.5% of men. These differences reflect, among other factors, the structure of labor markets and the distribution of administrative, information based, and cognitive occupations in which these systems can provide assistance.

The report nevertheless emphasizes that tasks are more likely to be transformed than entire occupations eliminated in many cases. A job may be exposed because some of its activities can be automated, while others still require experience, judgment, human interaction, professional accountability, or physical presence. Exposure is therefore an indicator of potential transformation, not a direct measure of how many positions will disappear.

This distinction is essential to understanding the effects on workers. An accountant may use generative tools to prepare documents or analyze data without the occupation itself disappearing. A teacher may use an assistant to prepare exercises, while continuing to perform work that depends on pedagogy, guidance, and understanding students. A journalist may automate certain research or data processing tasks while remaining responsible for verification, the prioritization of information, and editorial decisions. Conversely, when a position consists largely of standardized tasks and automation offers substantial gains, employers may reduce staffing, change the qualifications they seek, or fundamentally reorganize work. The effects therefore vary by occupation, company, country, and management decisions.

The ILO and NASK report also identifies differences by national income level and by the distribution of occupations between women and men. In high income economies, administrative and certain office based occupations, in which women are often strongly represented, show substantial exposure. The most exposed categories are therefore not distributed evenly across social groups. The eventual consequences, however, will also depend on access to training, companies’ ability to reorganize their activities, demand for new services, and social protection policies. Exposure figures should be used to prepare for the transition and identify training needs, not to announce the inevitable disappearance of human work.

14. The World Economic Forum: Job Creation, Displacement, and Occupational Restructuring

The World Economic Forum’s Future of Jobs Report 2025 estimates that major economic and technological transformations could contribute to the creation of approximately 170 million jobs and the displacement of 92 million others by 2030, producing a projected net increase of 78 million jobs worldwide. These figures are based on a survey of 1,043 employers across 55 economies and 22 industries, reflecting their expectations about changes in their operations and labor markets. They are not a count of jobs already created or eliminated. The survey provides useful insight into the intentions of participating employers, but it does not comprehensively represent all small businesses, informal employment, or working conditions in economies where data are less complete. It presents a projection that depends on its assumptions and on the decisions businesses, governments, and workers make in the coming years.

It is also important to emphasize that the World Economic Forum does not attribute these developments to artificial intelligence alone. Demographic changes, the energy transition, geoeconomic tensions, the cost of living, healthcare needs, supply chain transformations, and the adoption of other technologies all contribute to occupational restructuring. Jobs in healthcare, education, construction, logistics, and several essential services may grow for reasons that are not directly related to artificial intelligence. Conversely, certain administrative or repetitive functions may be affected by automation, the digitization of procedures, or corporate reorganization. Artificial intelligence is therefore a major factor, but it operates within a broader economic transformation.

The report’s projected global net increase does not mean that every country, region, or category of worker will experience a positive outcome. A job created in one sector does not automatically replace a job eliminated in another. New positions may require different skills, training, or geographic locations from those available to displaced workers. Someone whose job disappears may struggle to access emerging positions even if total employment increases. Global net job growth can therefore coexist with local crises, periods of unemployment, declining incomes, or widening inequality.

15. Goldman Sachs and the Macroeconomic Potential of Artificial Intelligence

In April 2023, Goldman Sachs Research estimated that generative artificial intelligence could increase global gross domestic product by approximately 7% over a period of about ten years, representing the potential for several trillion dollars in additional economic output. The same analysis discussed exposure affecting the equivalent of approximately 300 million full time jobs. These figures helped popularize the idea that the technology could bring about an economic transformation of exceptional scale. They must nevertheless be understood for what they are: estimates based on assumptions about technology adoption, the proportion of tasks that can be automated, and how productivity gains might spread across different sectors.

A job’s exposure to artificial intelligence does not mean that the position will be eliminated, or even that all the working time associated with it will be automated. An occupation may include tasks that can readily be assisted by a generative system alongside others that remain difficult to automate. Adoption may also be constrained by integration costs, regulatory obligations, error risks, supervision requirements, and organizational resistance. Productivity gains may emerge gradually as businesses adapt their procedures and invest in the necessary skills. Macroeconomic effects therefore depend as much on work organization and institutional conditions as on technical capabilities.

Even when productivity gains are realized, their consequences for employment and wages are not predetermined. A company may use the gains to expand its activities and create new positions, or it may seek to reduce costs by cutting its workforce. It may lower prices, improve its products, or increase its margins. Across an economy, higher productivity can support growth and generate new demand, but it can also concentrate benefits if markets are dominated by a small number of players. Goldman Sachs’ estimates should therefore be read as scenarios of economic potential, not as a promise of certain growth or a forecast of a precise number of job losses.

For countries with limited capabilities, the question is how to access productivity gains without increasing dependence excessively. Small businesses may benefit from affordable tools that improve certain tasks, but they may also find themselves at a disadvantage if competitors have better infrastructure, data, and expertise. Public administrations can improve their services, but only if they have reliable organizational systems and oversight procedures. Potential growth is not an automatic result of tool availability. It depends on the ability to integrate these tools into productive activities and distribute benefits beyond the businesses that are already best equipped.

16. McKinsey: Automation and Occupational Transitions

The McKinsey Global Institute’s 2017 report, Jobs Lost, Jobs Gained, outlined several automation scenarios extending to 2030. Depending on the scenario, between 400 million and 800 million people worldwide could be displaced by automation and need to find new employment. The report separately estimated that approximately 75 million to 375 million workers might need to change occupational categories and acquire new skills, depending on the assumed pace of adoption. These figures concern automation broadly, not generative artificial intelligence specifically. The current wave of generative AI had not yet reached the scale or form that we see today. The estimates should be read as conditional scenarios based on assumptions about technical feasibility, costs, and the speed at which technologies spread, rather than as certain predictions.

The report highlighted the importance of tasks in assessing automation. Occupations are not indivisible blocks: they comprise activities whose technical feasibility varies. A machine or software system may take over certain operations without being able to perform the entire job. This approach helps explain why an occupation can be transformed profoundly without disappearing. It also shows why two people with the same job title may experience different effects: their work organization, responsibilities, and available tools may differ considerably.

The transition depends on more than technology. It is also shaped by investment costs, regulation, infrastructure availability, social acceptance, and expected economic returns. Some companies may automate a function rapidly when procedures are standardized and the benefits can be measured immediately. Others may retain a more traditional organization because integration is too expensive, risks are high, or customers place considerable value on human involvement. Long term projections are therefore sensitive to economic assumptions and should not be treated as certain descriptions of what will happen in every sector.

The lasting lesson from this research is that occupational transitions must be anticipated before job displacement becomes widespread in the sectors concerned. Training policies should not be limited to teaching people how to use a handful of applications. They must strengthen transferable skills, data literacy, analytical thinking, communication, problem solving, and the ability to work alongside automated systems. In countries where retraining programs are weak, workers risk bearing a disproportionate share of transition costs even when the overall economic gains are substantial.

17. PwC: Productivity, Skills, and Wages

PwC’s Global AI Jobs Barometer 2026, published in June 2026, offers another perspective on the relationship between artificial intelligence, productivity, and skills. According to the indicators presented in this edition, companies operating in sectors more exposed to artificial intelligence recorded productivity growth of 34% in 2025 compared with 2018, versus 24% among companies PwC classifies as less able to use AI. The report also records employment growth of 52% in the first group, compared with 36% in the second, and wage growth of 24% versus 17%.

These comparisons describe changes observed across groups of companies and do not, by themselves, demonstrate that AI explains the entire difference. The report also identifies an average wage premium of 62% associated with AI skills in the data analyzed, up from 57% the previous year. PwC further reports that the top 20% of performers among companies highly exposed to AI achieved average labor productivity growth of 163% compared with 2018. These figures must be understood in light of PwC’s methodology and the populations it studied. They do not mean that every company adopting an AI tool will achieve similar results, or that the observed differences can be attributed to a single cause. The sectors, businesses, and workers best positioned to adopt these technologies may already possess structural advantages.

The figures must be interpreted with attention to the methods used, the populations observed, and the definition of the skills concerned. A wage premium associated with AI skills does not mean that everyone who completes a training course will automatically receive a comparable raise. It may reflect strong demand for scarce expertise, differences in responsibilities, or the concentration of these skills in companies able to offer higher salaries. Similarly, productivity growth in exposed sectors may result from a combination of factors, including investment, work organization, demand, and changing business models. The observed associations help explain emerging trends, but they should not be treated as universal proof of causation.

The report nevertheless highlights an important issue: the skills employers seek may evolve faster than traditional education and training pathways. Companies may expect new employees to have more advanced digital skills, stronger analytical abilities, and the capacity to verify automatically generated results. Some entry level positions could be particularly affected if tasks once assigned to junior employees are progressively automated or delegated to assistance systems. This outcome is neither uniform nor inevitable, but it raises a question for organizations: how can future professionals gain experience if some of the tasks that previously helped them develop it disappear or change in nature?

Educational institutions, businesses, and governments will therefore need to design pathways that combine foundational knowledge, practical experience, mentoring, and continuous learning. The goal should not be to turn every worker into an artificial intelligence specialist, but to enable people to understand the tools relevant to their profession, recognize their limitations, and retain their capacity for independent judgment. For developing countries, access to these skills can become a significant advantage if training is accessible, adapted to local needs, and connected to real employment opportunities. Otherwise, wage premiums and productivity gains risk becoming concentrated in the businesses and economies that are already best equipped.

18. The International Monetary Fund: Uneven Global Exposure

In January 2024, the International Monetary Fund estimated that approximately 40% of global employment was exposed to artificial intelligence, including close to 60% in advanced economies, 40% in emerging market economies, and 26% in low income countries. These figures are based on a broad approach to occupational exposure to artificial intelligence and are not directly comparable with those of the ILO and NASK, which focus on generative AI and use a specific methodology. They nevertheless confirm that the potential effects of the technology are distributed unevenly according to economic structure and the composition of employment.

In advanced economies, the large proportion of administrative, analytical, and cognitive occupations may increase exposure to certain AI applications. These economies, however, generally have greater access to capital, infrastructure, and training systems to support adoption. Low income countries may have lower direct occupational exposure because of the greater role played by certain physical or informal activities, but this does not protect them from the consequences of technological change. They may experience indirect effects through changes in trade, services, investment, and international labor markets while lacking the resources needed to benefit from new tools.

The IMF’s analysis therefore invites a distinction between two dimensions. The first is the vulnerability of existing jobs to automation or task transformation. The second is an economy’s ability to benefit from productivity gains, attract investment, and train its workforce. A country may be relatively less exposed to direct automation yet still be disadvantaged if its businesses lack access to the necessary infrastructure and skills. Conversely, a highly exposed economy may achieve substantial productivity gains if it organizes adoption effectively, protects workers, and encourages the emergence of new activities.

For governments, the challenge is to develop policies that reflect their productive structures and institutional capabilities. The appropriate response cannot be identical in an advanced economy with major technology companies and in a small island state facing infrastructure deficits. Governments must identify the sectors in which gains are most attainable, strengthen priority skills, improve the business environment, and establish suitable protection mechanisms. A global average does not provide a national strategy. It signals the need for analysis tailored to each context.

19. The Digital Divide Could Turn Artificial Intelligence into a Driver of Inequality

Access to artificial intelligence is often presented as a universal opportunity, but the benefits depend on the resources available to users. A business with access to high quality data, reliable connectivity, trained employees, and investment capacity can integrate new tools more quickly. A small business without these resources may remain limited to occasional use, without achieving comparable productivity gains. Infrastructure and skills gaps may therefore widen even when applications are publicly accessible. The digital divide is no longer just about whether people can connect to the Internet; it is about whether they can turn connectivity into economic and social value.

Research on development gaps and artificial intelligence, including analyses published in 2026 by international institutions on differences in occupational tasks and inequalities in digital access, underscores the importance of these asymmetries. The specific findings of each study must nevertheless be verified, and no report should be credited with conclusions broader than those it actually establishes. The general principle remains clear: economies with limited resources may lack the infrastructure, skills, and institutions needed to turn innovation rapidly into productivity gains. This can affect their competitiveness, their ability to create skilled jobs, and their position in international value chains.

The language divide adds another dimension to this inequality. Languages with extensive digital corpora, research resources, and large markets are often better represented in generative tools. Languages with fewer resources may receive less effective support, particularly in fields where accuracy, cultural references, and contextual understanding are essential. For Haiti, the quality of resources in Haitian Creole is therefore a matter of economic participation and cultural sovereignty. Institutions, universities, media organizations, and civil society groups can help build reliable language resources while respecting individual rights and data protection principles.

Reducing these gaps requires a strategy that goes beyond simply distributing tools. It requires investment in connectivity, electricity, education, digital skills, high quality data, and institutional capacity to evaluate technology. International partnerships can help, but they should be designed to strengthen local capabilities rather than establish permanent dependence. The goal is not to reproduce the same infrastructure everywhere on the same scale. It is to give countries the means to participate in the digital economy, negotiate the terms of access, and retain a reasonable share of the value created.

20. What Statistics Cannot Decide for Us

The reports produced by the ILO, the World Economic Forum, Goldman Sachs, McKinsey, PwC, and the IMF provide valuable information about possible changes in the world of work. They do not, however, constitute a single verdict on the future of employment. These studies do not all examine the same technology, cover the same periods, or define exposure in the same way. Some assess tasks that may be assisted or automated; others model job creation and displacement, while still others examine observed relationships between skills, productivity, and wages. Their value lies in identifying trends, risks, and scenarios, provided that the methodological limitations of each study are respected.

Technical capability must also be distinguished from economic consequences. The fact that a system can perform a task does not mean that a company will adopt it immediately. Integration costs, output quality, security requirements, legal obligations, and user trust may slow or limit its use. Even when a task is automated, the corresponding job may evolve rather than disappear, as new responsibilities emerge or demand increases. Conversely, partial automation may be sufficient to reduce staffing needs in certain organizations, especially when activities are standardized and financial margins are under pressure.

The distribution of gains matters as much as their scale. Productivity growth can raise incomes, lower prices, and create new activities, but it can also increase the margins of a small number of companies or weaken the bargaining power of certain workers. The consequences depend on competition, training, collective bargaining, public policy, and the opportunities available to displaced workers. An economy can register positive growth while working conditions deteriorate for certain groups. The success of the transition must therefore be assessed using several indicators, including productivity, job quality, incomes, access to training, inequality, and institutional capacity to manage change.

For developing countries, this methodological caution is particularly important. Global estimates often rely on occupational categories and datasets that do not fully reflect the importance of informal work, linguistic specificities, or the constraints facing small economies. National decisions should draw on international research while also producing analyses adapted to local realities. Haiti and the Caribbean need reliable data on their own sectors, training needs, infrastructure, and adoption opportunities. Without such knowledge, they risk replicating strategies designed for economies with very different resources and priorities.

21. Preparing for the Future: From Technology Adoption to the Capacity to Act

Artificial intelligence will not eliminate the need for human decisions; it will change the conditions under which they are made. Governments will need to select priority sectors, establish rules for data use, and protect essential services. Businesses will need to assess expected benefits, train employees, and determine which functions can be automated without compromising the quality of their products or services. Educational institutions will need to adapt their curricula without abandoning foundational knowledge, critical thinking, or the ability to verify information. Media organizations will need to learn how to use research and analytical tools while preserving independent verification and editorial accountability.

Preparation must begin with skills development. Learning how to formulate prompts for a generative system is not enough. Users must understand the limitations of these tools, recognize plausible errors, protect confidential information, and know when human intervention is indispensable. Training must be adapted to different occupations and starting skill levels so that artificial intelligence does not become another mechanism of exclusion for those who have had the fewest opportunities to access digital education. Experienced workers must be able to update their skills, while young people need a sufficiently strong educational foundation to adapt as tools and practices change.

The second priority is institutional. Organizations adopting artificial intelligence systems must define responsibilities, regulate data access, document their uses, and establish oversight mechanisms. In public administration, decisions affecting people’s rights must be subject to appropriate supervision and, where necessary, avenues of appeal. In business, contracts with providers should be assessed in terms of security, confidentiality, long term costs, and the possibility of switching solutions. Governance is not an external obstacle to innovation. It helps reduce risks, strengthen trust, and make benefits more sustainable.

The third priority concerns infrastructure and cooperation. Countries with limited resources may achieve better results by focusing investment on essential needs, pooling selected capabilities, and developing partnerships that strengthen local expertise. They must compare the full costs of solutions, plan for maintenance and training, and avoid projects whose recurring expenses would remain beyond their financial capacity. Regional cooperation can support education, research, cybersecurity, language resources, and disaster preparedness. It can also strengthen negotiating power with international providers. Access to a tool alone does not constitute a development policy. Countries must also have the means to use it reliably, assess its results, and adapt it to the needs of their populations.

Finally, the transformation must be assessed in terms of its effects on society. Success cannot be measured solely by the number of models in use, the computing power of data centers, or the volume of announced investment. It must also be reflected in improved services, more equitable access to skills, quality employment, greater resilience, and an enhanced ability to make informed decisions. Artificial intelligence can contribute to these objectives, but it will not achieve them automatically. The outcome will depend on the priorities chosen and on the determination to make technology a means of development rather than an end in itself.

ARTIFICIAL INTELLIGENCE, THE NEW FRONTIER OF SOVEREIGNTY

Artificial intelligence is ushering in a period of transformation whose consequences extend far beyond software and digital innovation. It is changing the conditions of production, relationships between businesses and workers, the organization of scientific research, communications infrastructure, and the balance of power between states. It is progressively influencing how knowledge is created, decisions are prepared, resources are mobilized, and wealth is distributed. In this emerging global landscape, the United States benefits from a particularly integrated technology ecosystem combining capital, companies, research, and infrastructure. China is investing to strengthen its domestic capabilities and reduce its strategic dependencies. Taiwan retains a pivotal position in advanced semiconductor manufacturing, while Europe is seeking to translate its scientific, industrial, and regulatory strengths into greater strategic autonomy. Large-scale industrial projects, investments in data centers, the race to develop specialized processors, and competition over satellite networks all point to a profound shift: mastery of artificial intelligence increasingly depends on control over the physical, energy, financial, and human infrastructure that makes its development possible.

Yet this transformation cannot be understood solely through the competition between major powers. Its consequences are also evident in companies reorganizing their operations, workers adapting their skills, educational institutions rethinking their methods, and public administrations seeking to improve their services. Employment projections confirm the scale of the changes that may lie ahead, but they do not make it possible to predict a uniform future. A profession’s exposure to artificial intelligence does not necessarily mean that the profession will disappear, just as a productivity gain does not guarantee a fair increase in wages or a general improvement in working conditions. Net job creation worldwide does not automatically protect the people, regions, and sectors facing occupational displacement. International studies are valuable tools for anticipating change, provided that their methodologies, observation periods, and underlying assumptions are taken into account. Their real purpose is to inform decisions and prepare societies for transitions, not to present scenarios as inevitable outcomes.

For Haiti, the Caribbean, and small states, the challenge is neither to reproduce immediately the industrial capabilities of major powers nor to claim mastery over every available technology. It is to identify realistically the areas in which artificial intelligence can address priority needs, the infrastructure that must be strengthened, the skills that must be developed, and the dependencies that need to be managed. Education, agriculture, healthcare, disaster risk management, public services, research, the media, and the development of linguistic resources all offer areas in which targeted investment could strengthen national capabilities and improve development prospects. These opportunities, however, will produce lasting results only if they rest on credible institutions, reliable data, appropriate infrastructure, effective human oversight, rigorous protection of rights, and transparent cost assessments. Without these foundations, artificial intelligence risks remaining an imported technology whose benefits are limited, whose applications are insufficiently adapted to local realities, and whose conditions of access are largely determined by external actors.

Digital sovereignty does not mean that every country must design, manufacture, and finance everything within its own borders. Such an ambition would often be unrealistic, particularly for economies with limited resources. Rather, it requires sufficient capacity to understand the systems being used, assess their limitations, choose partners, protect data, train citizens, and retain meaningful decision-making options when economic or geopolitical circumstances change. For small states, this autonomy can be strengthened through regional cooperation, shared infrastructure, university partnerships, scientific training, and investment in human capabilities. It also requires technology contracts, equipment choices, and data policies to be examined in light of their long-term consequences. Sovereignty is therefore measured not merely by the possession of a technology, but by the ability to use it in the collective interest without relinquishing control over essential decisions.

This requirement brings the human being back to the center of the debate. Computing power, execution speed, and productivity gains are not, in themselves, sufficient indicators of progress. Technological transformation must also be assessed in terms of its contribution to education, dignified work, reduced inequality, the quality of public services, individual freedom, and people’s ability to participate in economic and social life. Businesses play a decisive role in creating and adopting new capabilities, but governments, educational institutions, the media, and civil society must also help define the rules, safeguards, and collective priorities. The objective is not to prevent innovation, but to ensure that its benefits are not reserved for those who already possess the greatest capital, infrastructure, and influence.

The central question of our time is therefore not simply whether machines will become more capable. Their progress is expected to create new opportunities across many fields, although that progress will remain uneven and subject to scientific, industrial, energy, and economic constraints. The real question is whether societies can organize this power, prevent its abuse, support workers through changing conditions, and distribute its benefits more equitably. Artificial intelligence can deepen inequalities or expand human possibilities; it can reinforce the concentration of power or encourage new forms of cooperation; it can increase societies’ dependence or contribute to their autonomy. None of these outcomes is guaranteed in advance. They will depend on industrial choices, public policy, the quality of education, the strength of institutions, and citizens’ ability to participate in decisions that shape their future.

For CPGNNews, this reflection falls squarely within the responsibilities of an international public-interest media organization. Examining artificial intelligence means analyzing a transformation that simultaneously affects the economic, political, social, scientific, technological, and cultural dimensions of the contemporary world. It also means looking beyond spectacular announcements to examine power relations, verify claims, distinguish established facts from projections, and give a voice to societies whose realities are too often marginalized in dominant narratives about technology. From Haiti and the Caribbean, CPGNNews seeks to contribute to a broader understanding of these global developments by connecting decisions made in major centers of power with their potential consequences for peoples, economies, and future generations. Born in the Caribbean, Open to the World! expresses more than a geographical identity. It embodies an editorial ambition: to begin from a regional foundation while illuminating universal issues, without disregarding the diversity of national experiences or compromising independent judgment. The Caribbean is our point of origin, not our limit.

Ultimately, artificial intelligence confronts us with a responsibility that extends beyond businesses, governments, and national borders. It compels us to consider the society we want to build, the values we wish to preserve, and the opportunities we intend to pass on to future generations. In a world where technological power is becoming a major instrument of competitiveness, security, and influence, sovereignty cannot be reduced to the ability to manufacture more advanced machines. It must also mean that peoples can understand the transformations affecting them, participate in the decisions guiding those transformations, and defend their interests without abandoning international cooperation. True progress will therefore be measured not only by what artificial intelligence makes possible, but also by how humanity chooses to distribute those opportunities, assume the associated responsibilities, and limit the most damaging consequences.

The future will not be determined solely by those who build the most powerful machines, but also by those who know how to place that power at the service of human dignity, justice, and shared development. In the age of unbridled intelligence, the fundamental challenge is not to give machines ever greater power, but to ensure that this power helps make humanity freer, more responsible, and better prepared to shape its common destiny.

DOCUMENTARY AND INSTITUTIONAL REFERENCES

The following references constitute the documentary foundation of this report. Forward-looking figures should be interpreted in accordance with the methodologies and assumptions specific to each publication. Industrial announcements are distinguished from capacities that are already operational, and press reports are attributed to their respective sources.

1. International Labour Organization (ILO) and NASK: Generative AI and Jobs: A Refined Global Index of Occupational Exposure, May 2025. https://www.ilo.org/publications/generative-ai-and-jobs-refined-global-index-occupational-exposure

2. World Economic Forum (WEF): The Future of Jobs Report 2025. https://reports.weforum.org/docs/WEF_Future_of_Jobs_Report_2025.pdf

3. Goldman Sachs Research: Generative AI Could Raise Global GDP by 7 Percent, April 5, 2023. https://www.goldmansachs.com/insights/articles/generative-ai-could-raise-global-gdp-by-7-percent.html

4. McKinsey Global Institute: Jobs Lost, Jobs Gained: What the Future of Work Will Mean for Jobs, Skills, and Wages, 2017. https://www.mckinsey.com/featured-insights/future-of-work/jobs-lost-jobs-gained-what-the-future-of-work-will-mean-for-jobs-skills-and-wages

5. PricewaterhouseCoopers (PwC): Global AI Jobs Barometer 2026, June 2026. https://www.pwc.com/gx/en/issues/artificial-intelligence/publications/artificial-intelligence-study.html

6. International Monetary Fund (IMF): Kristalina Georgieva, AI Will Transform the Global Economy. Let’s Make Sure It Benefits Humanity, January 2024. https://www.imf.org/en/Blogs/Articles/2024/01/14/ai-will-transform-the-global-economy-lets-make-sure-it-benefits-humanity

7. International Labour Organization and World Bank: Disruption without dividend? How the digital divide and task differences split GenAI’s global impact, joint working paper, 2026. The study covers 135 countries and examines how gaps in digital infrastructure and differences in task composition may cause developing economies to experience disruption before realizing the anticipated productivity gains. https://www.ilo.org/publications/disruption-without-dividend-how-digital-divide-and-task-differences-split

8. European Commission: Strengthening Europe’s Tech Sovereignty, June 3, 2026. Presentation of the European package addressing semiconductors, artificial intelligence, cloud computing, open-source software, and the digitalization of energy systems. This is an announced program and a set of policy proposals, not evidence that European technological autonomy has already been achieved. https://commission.europa.eu/news-and-media/news/strengthening-europes-tech-sovereignty-2026-06-03_en

9. European Commission: Communication on European Tech Sovereignty, accompanied by an EU Open Source Strategy, June 3, 2026. A public policy document detailing announced initiatives relating to semiconductors, cloud and AI infrastructure, open-source software, and energy. https://digital-strategy.ec.europa.eu/en/library/communication-european-tech-sovereignty-accompanied-eu-open-source-strategy

10. North Atlantic Treaty Organization (NATO): The Ankara Summit Declaration, July 8, 2026. https://www.ncia.nato.int/newsroom/news/the-ankara-summit-declaration

11. SpaceX: Breaking Ground on Terafab in Texas, industrial project announcement, August 6, 2026. The source specifies that the terawatt of computing power refers to the combined anticipated demand of Tesla and SpaceX, rather than to an announced production capacity for the facility. https://www.spacex.com/updates/starship-2025-report

12. Reuters: SpaceX, Tesla to initially spend $16.8 billion on Terafab chip plant in Texas, August 6, 2026. A complementary journalistic source on the announced initial investment and the proposed manufacturing facility. https://www.reuters.com/business/media-telecom/spacex-says-terafab-be-built-texas-with-initial-investment-168-billion-2026-08-06/

13. Government of Texas: Announcement concerning SpaceX’s expansion in Grimes County, August 6, 2026. https://gov.texas.gov/news/post/governor-abbott-announces-spacex-expansion-in-grimes-county

14. Reuters: Russia Plans to Launch Its Smaller Version of Starlink Next Year, June 5, 2026. https://www.internazionale.it/ultime-notizie-reuters/2026/06/05/russia-plans-to-launch-its-smaller-version-of-starlink-next-year

15. Reuters: Russian Satellite Activity Raises Concerns for Starlink, German Space Commander Says, October 9, 2026. https://www.reuters.com/business/media-telecom/russian-satellite-activity-raises-concerns-starlink-german-space-commander-says-2026-10-09/

16. Middle East Institute: AI, the Gulf, and the US: A Primer, February 2026. https://mei.edu/report/ai-the-gulf-and-the-us-a-primer/

17. OpenAI: The Next Phase of the Microsoft–OpenAI Partnership, April 27, 2026. https://openai.com/index/next-phase-of-microsoft-partnership/

18. ASML: Institutional documentation on lithography systems used in semiconductor manufacturing. This source provides context for the role of the Dutch company in the semiconductor industry’s value chain. https://www.asml.com/en

METHODOLOGICAL NOTE

The forward-looking figures presented in this report are not directly interchangeable. The ILO and NASK studies examine exposure to generative artificial intelligence; the IMF estimates use a broader definition of exposure to AI; the World Economic Forum reports the expectations of surveyed employers; Goldman Sachs and McKinsey present economic scenarios based on explicit assumptions; and PwC analyzes observed associations between AI exposure, productivity, employment, and remuneration.

Industrial announcements, particularly those concerning Terafab and satellite constellations, are presented as projects or intentions whenever their full implementation has not been established. Data and projects will need to be updated as developments unfold, without turning projections into accomplished results.

CPGNNews — Independent International Public Interest Media │ Research • Investigation • Analysis • Journalism

Born in the Caribbean, Open to the World!

By Yvon BONHOMME
Independent Researcher, Essayist, Geopolitical Analyst and Governance Specialist
Editor-in-Chief, CPGNNews

Email: info@cpgnnews.com │ © All rights reserved.