How AI is reshaping strategic foresight and global power

Artificial intelligence is changing how we think about the future and who gets to shape it. For strategic foresight – the discipline of mapping out possible futures to make better choices today – this moment presents both an opportunity and a test. AI can help practitioners scan more widely, identify emerging signals and trends, and explore possible futures with new speed and scale. But foresight has never been only about producing scenarios. Its real value lies in helping leaders and policy-makers question assumptions, prepare for uncertainty and build strategic decisions that ensure long-term resilience.

As AI rapidly reshapes economies, geopolitics, institutions and societies, foresight practitioners must ask a deeper question: How do we move beyond mere adoption to truly responsible use? The challenge is to integrate AI in ways that strengthen human judgement rather than replace it, ensuring that faster analysis does not come at the cost of deeper thinking, contestation and strategic imagination.

For much of modern history, power in international relations was easy to recognize. It belonged to countries with armies, aid budgets, reserve currencies, diplomatic alliances and seats in the world’s major institutions. AI is changing that. It is becoming one of the clearest markers of power in our post-hegemonic era. The countries and companies that control compute, chips, models, cloud infrastructure and talent will increasingly shape who gets heard, who gets funded and who gets to build the future.

AI is becoming a new source of strategic power because it depends on more than models alone. It requires chips, cloud infrastructure, data centres, energy, talent, regulation and trust. This is why recent AI infrastructure projects, such as the UAE-US Stargate announcement, matter beyond their technical details. They point to what is sometimes referred to as “infrastructure diplomacy for the AI age”: a world in which data centres and cloud regions begin to resemble ports, pipelines or other strategic assets. The old markers of power will not disappear – armies, capital and diplomacy still matter. But they are joined by a new currency of power: compute.

This shift is creating new openings for middle powers, who are leveraging these infrastructure needs to carve out global influence. For example, the UAE is using AI infrastructure to position itself as a connector for the US, the Gulf, Africa and Asia. And Singapore has announced more than S$1 billion in public AI research funding through 2030.

This is also where AI begins to look like the new foreign aid. As traditional aid budgets contract – OECD preliminary data shows official development assistance fell 23.1% in 2025, the largest annual contraction on record – AI infrastructure and AI-enabled services are becoming a new channel of development influence. The country that provides the cloud region, the AI tutor, the health model or the government chatbot gains the kind of influence once associated with aid donors. AI diplomacy is not just about models. It is about electricity grids, water, debt, sovereignty, contracts and who ultimately controls the infrastructure. The future of development may be less about donor conferences and more about data centres.

The third shift is perhaps the strangest: some AI and big technology firms are beginning to resemble quasi-sovereign actors. They aren’t countries but they increasingly act as non-state powers because they control key things that states need. Microsoft’s $1.5 billion investment in G42 included a minority stake, a board seat and security assurances to the US and UAE governments. That does not make Microsoft a state. But it does show how frontier firms are becoming strategic intermediaries between states.

In addition to changing which actors shape international relations, AI is also changing how those actors think. When governments rely on AI to scan information, summarize events and generate scenarios, those systems begin to shape what filters up to leaders and what is missed. As such, AI influences which futures feel plausible, which threats feel urgent and which options appear on the table. So, while AI doesn’t have a seat at the table, it is starting to set the agenda.

The future of international relations will be driven by a new cast of characters – the legacy powers of sovereign states and the international order, as well as new players such as tech giants and technology itself. The leadership question is no longer whether to accept this but how to build the foresight, institutions and imagination to govern with it.

If AI is reshaping the world that foresight seeks to understand, it is also changing the methods practitioners use to understand it. When treated not as an oracle but as a powerful, fallible research instrument, AI dramatically expands the research capacity of foresight teams.

There are four primary ways in which AI can be integrated into the foresight workflow:

AI can also accelerate synthesis and scenario development. It can help identify key factors, map relationships between drivers, test assumptions and generate multiple coherent narratives at speed. These are what have previously been referred to as “first-round futures”: early maps of the possibility space that can support deeper human deliberation. But it’s important to treat such outputs as drafts, not conclusions. AI can provide structure, speed and breadth, but foresight teams must still define the question, set quality criteria, validate sources and decide what is strategically meaningful. In this sense, AI’s promise is not to replace foresight expertise, but to give practitioners a richer and more expansive research base from which human judgement can begin.

The same qualities that make AI valuable for foresight are also what makes it risky. The risks of AI are not only technical: they are cognitive, institutional and political. AI can hallucinate facts, reproduce dominant narratives, compress deliberation and generate results that look more convincing than they are.

Across corporate strategy teams, governments and international organizations, generative AI is already reshaping how early-stage foresight is conducted, but speed comes with a paradox. The faster foresight outputs are generated, the greater the risk that we stop building the thinking behind them. When AI-generated outputs are treated as deliverables rather than starting points, the deliberative work that gives scenarios their strategic value – the disagreement, the reframing, the discomfort – gets compressed or skipped entirely.

Speed can also create a more subtle illusion. As AI systems process vast amounts of data and generate coherent narratives, they can give the impression that the future is becoming more predictable. This is perhaps the sharpest risk: foresight properly practised does not aim to predict the future; it aims to expand the range of futures we can imagine and prepare for. The challenge is not whether to use AI, but how to do so without hollowing out the practice.

To avoid these pitfalls, the following four principles are worth holding onto.

AI can be a very useful collaborator in a scenario process, but humans remain essential because scenarios are not only descriptions of possible futures: they are about changing how people perceive uncertainty, argue about choices and decide what to do next. We should, therefore, be intentional about what could happen fast, with more AI, and what should happen slow, with more humans.

AI appears to distribute foresight power to non-experts. What it can actually do, however, is concentrate epistemic authority invisibly and make the resulting conclusion unchangeable.

When a small group controls how questions are framed, AI does not merely reflect their assumptions: it amplifies and legitimizes them, producing outputs that feel analytical and comprehensive precisely because they are so fluently structured. Accountability, which should rest with identifiable human judgement, quietly diffuses across the tool itself. In a world where there is no named person or organization positioned to contest, the conclusions become, in practice, unchallengeable: not because they are correct, but because there is nowhere to push back against. This is foresight capture, and it is all the more seductive because whoever owns the prompting ultimately lays claim to how we understand the future.

The question leaders should be asking is not whether their AI tools are sophisticated enough, but whether their foresight process is one that people inside the organization can actually engage with, challenge and therefore trust.

This point is sharpened by the tension between access and signal. AI may make foresight tools more widely available, but more people producing scenarios does not automatically mean better foresight. If organizations rely on the same models, prompts and dominant assumptions, AI can widen participation while narrowing imagination – returning familiar worldviews in polished, plausible form.

Consequently, arguments such as the call for intergenerational foresight are not simply about representation. They’re about signal quality. Those who will live longest with today’s decisions bring a relationship to the future that is less shaped by sunk costs and existing institutional commitments, while older leaders bring experience, memory and a deeper relationship to consequences. The value lies in the dialogue between them. AI can generate outputs quickly, but it cannot manufacture the trust, mentorship and transfer of tacit knowledge that make such dialogue meaningful.

As co-dependency on AI increases, the importance of depending on a hybrid model that uses the research capabilities of agentic AI while maintaining human judgement as priority is increasing. The following three rules can drive hybrid intelligence:

Before prompting, we need to establish our intent. This means knowing what question we are answering, for whom, by when, and what kind of content will be most useful.

To be effective, this intent must define three concrete components:

The trap is that what looks “comprehensive” and “well-structured,” which AI delivers by default, is not necessarily good quality. Being specific about data requirements, minimum triangulation and sources to use all drive quality. Writing a brief before opening the AI tool can boost answer quality dramatically. Asking the AI tool to provide a template for a good prompt is another powerful practice.

Every AI-generated artefact should carry an implicit label: draft, until validated.

Three checks matter most:

The trap to avoid is formatting credibility. A neatly structured profile reads as rigorous even when the reasoning is thin. Structure must be treated as a useful default, not as evidence of quality.

The third activity is the least discussed and the most consequential. Throughput is the process by which AI outputs are generated. Three practices help to drive analytic quality during this phase:

The challenge of hybrid intelligence does not end with better prompting or stronger validation. AI usage in foresight can also change the leadership conditions around the practice: who frames the questions, who accepts the answers, who challenges the assumptions, and who ultimately owns the scenario of the future.

AI’s impact on the field of strategic foresight and future thinking is not a pure good or evil. The technology has the ability to facilitate some aspects of scenario creation and data analysis but also carries with it the potential to replace thoughtful human analysis with output that is polished, yet generic and lacking insight.

The challenge for foresight practitioners is therefore neither to reject AI integration nor to embrace it uncritically, but to critically assess the evolving role of AI systems and technology companies as influential non-state actors in global politics and future planning.

Human judgement must remain integral to the foresight process, working in tandem with AI to develop more plausible, nuanced and responsible futures. At a time when AI-driven disruption is making the future increasingly complex and uncertain, strategic foresight has become more important than ever. By balancing the opportunities offered by AI with an awareness of its limitations, foresight can continue to support leaders in preparing for a broad range of possible futures.