A survey of 167 strategic foresight experts from 55 countries, conducted by the World Economic Forum together with the OECD, shows that artificial intelligence is already changing the practice of working with the future, but is not replacing human thinking. Experts value AI above all for saving time on routine tasks, yet they express serious concerns about the quality and reliability of the content it generates.
The structural forces here are obvious: state and international institutions face the growing complexity of global risks — from climate change to geopolitical shifts. Under these conditions, strategic foresight becomes not a luxury but a necessity for anticipatory governance. Historically, such instruments, as in the case of the Club of Rome reports or the early OECD models, have always depended on a balance between data and interpretation. Today AI accelerates the collection and processing of data, but, as respondents note, it suffers from hallucinations and an inability to reason inductively.
The conjunctural moment underscores the urgency: many governments are already integrating AI into planning processes, yet without clear protocols. The hidden layer is the interest of technology companies and consultants who promote their solutions, while experts caution against overestimating the capabilities of algorithms. The real incentives converge on one point: everyone wants to obtain scenarios faster and more cheaply, but no one is ready to entrust the future entirely to machines.
The dominant convergence points to the emergence of hybrid models. AI will process large volumes of information and identify patterns, while people will provide context, ethical assessment and creative foresight. This is the most likely outcome in the coming 12–18 months, since institutional constraints and concerns about trust will not allow the process to be fully automated. The two strongest counterarguments — the accelerated development of models with better reasoning and budgetary pressure to cut staff — could shift the balance, but only if the quality of AI rises substantially, which is not yet being observed.
The key indicator to watch in the coming 4–8 weeks: the publication of the first pilot projects on the use of AI in national foresight strategies in OECD countries and decisions on including or excluding AI tools from official guidance on anticipatory governance. It is precisely these steps that will show whether the hybrid approach takes hold or skepticism prevails.
Governments should already be developing internal standards for verifying AI outputs, so as not to lose control over strategic decisions.
