How is AI changing public affairs?

Public affairs is often imagined as a process in which leading experts from different fields meet with decision-makers, bringing research, impact assessments and data to the table. The well-informed decision-maker then carefully weighs the alternatives and arrives at the best solution. But this picture overlooks one important fact: decision-makers are human, too.

They work under pressure, make decisions on dozens of separate issues at the same time, and receive enormous amounts of information from people who all have good reasons to advocate for their own cause.

AI is making this information environment even more intense. When producing information becomes effortless, decision-makers can instantly be presented with an endless stream of seemingly credible memos, analyses and counterarguments. Material generated by AI in seconds may take a human hours to unpack.

This information overload fuels an asymmetry: claims are cheap to produce, but costly to evaluate.

Influence can be understood through two types of cognitive processes (De Bruycker 2025). Slow influence targets deliberate and analytical decision-making, while fast influence operates more through intuition, social cues and relationships. In this sense, the psychological models of Daniel Kahneman and Robert Cialdini can also be applied to public affairs.

In slow influence, decision-makers are provided with political, legislative, economic, or technical information that helps them evaluate different alternatives. At its best, both sides devote time and thought to the issue.

Fast influence, by contrast, focuses on emotion, experience, and intuition. What matters is who is delivering the message: Is the person familiar and trusted? Who else agrees with them? Does the proposal appear mainstream or marginal? What kind of authority stands behind the message, and does cooperation feel safe?

Influence, like all human interaction, involves reciprocity, appeals to authority, personal preferences and social conformity. Although these principles may look like a collection of psychological tricks, it is important to understand that influence has always been social. Being right is rarely enough to ensure that you are heard.

As information overload grows, it would be easy to conclude that advocates should simply make their messages even shorter and replace lengthy reports with sharp slogans. But that is not enough. Image-building without credible substance is a poor strategy, especially in long-term relationships with decision-makers. Trust develops slowly but can be damaged by a single factual error.

In the age of AI, successful public affairs therefore requires both fast and slow influence. Advocates must be able to provide in-depth information when decisions are being prepared, bring the issue into the wider public debate, and make it understandable, timely and easy to assess so that the recipient does not first have to spend half a working day figuring out why it matters.

Effective public affairs in the AI era therefore requires a broader approach:

• From individual memos to information infrastructure
  A PDF sent to a decision-maker is no longer enough. It needs to be supported by a broader information infrastructure. The information provided should also be verifiable through public sources, presented in an understandable format and connected to relevant background information. An organisation’s public information footprint also affects what decision-makers and AI models can find about the issue.
• Advocates need to anticipate counterarguments
  Decision-makers can already ask AI to explain in five seconds why a proposal is a bad idea. Advocates therefore need to understand in advance the strongest counterarguments to their position, the most significant points of contention and the key uncertainties. The most credible advocate is one who proactively recognises counterarguments and is prepared to address them.
• Personal relationships become more important
  Visual polish and fluent writing increasingly tell us less about the quality of information, as producing high-quality-looking background materials requires less effort than before. What becomes decisive is whether the decision-maker knows the organization, whether the information it has previously provided has proved accurate, and whether the organization is willing to acknowledge facts that do not support its own interests. Decision-makers must also be able to pick up the phone and ask an expert what actually lies behind a particular figure or memo. AI makes personal trust more valuable than ever.
• An ecosystem of influence replaces individual meetings with decision-makers
  Results also depend on a long-term and systematic approach to influencing. Formal participation in legislative preparation, parliamentary committee hearings, statements and expert contributions all exist within a broader context shaped by public debate, the media, partners, and the way the issue is framed in society. If a decision-maker encounters the same idea in an expert statement, a media discussion, comments from a trusted colleague and within their own stakeholder network, its impact is stronger.