In conversation with psychologist Rakhi Rajani: Why fear, not strategy, is driving much of the AI boom

Companies are pouring billions into AI, and some of it is being spent out of fear. It’s enterprise FOMO, fear of missing out, and it drives some of the least strategic decisions in the boardroom.

To understand the behavioural science beneath it, we sat down with Nearform advisor, chartered psychologist and computer scientist, Rakhi Rajani.

Rakhi has spent her career at the sharp end of new technology, most recently as Chief Digital and Strategy Officer at Genomics England, and previously as Expert Associate Partner at QuantumBlack and McKinsey Digital Labs.

She now works with CTOs, AI leads and C-suite teams, and is an advisor to Nearform. We spoke to her about social proof, action bias, and what a psychologically healthy relationship with AI might actually look like.

A: FOMO is often the visible emotion, but there are several quite different psychological mechanisms sitting underneath it, and each one drives a different kind of behaviour. Too often the conversation starts with “what's our AI strategy?” when it should start with “what problem are we solving?”

A: When reliable information is scarce, we use other people's behaviour as evidence of what we should be doing. We're looking for social proof under uncertainty. And with AI right now, executives are seeing a highly distorted sample: the announcements, the successful pilots, the conference stories. Very few people talk about the failed experiments, the weak adoption or the disappointing returns. So you get what we call pluralistic ignorance: everyone is privately uncertain, and each person reads everyone else's public confidence as evidence that they're further ahead.

A: It can tip them into action bias. Faced with a large, ambiguous threat, activity restores a sense of agency. So organisations run pilots, form task forces, stand up AI councils and set adoption targets. Some of that is valuable experimentation, but the useful distinction is between activity designed to generate knowledge and activity designed to reduce anxiety. For every AI pilot I'd ask: what uncertainty is this designed to resolve, and what result would cause us to stop? If neither is clear, the pilot may be managing anxiety more than generating knowledge.

A: It becomes about identity, and for senior leaders that's very personal. Executive authority has traditionally rested on judgement, expertise and synthesis, and AI increasingly operates in exactly those domains. That isn't simply a fear of being replaced. The more interesting question is where uniquely valuable human judgement now sits, and there's a status dimension too: how will I be perceived by my peers? “We automated our back-end operations” doesn't sound as impressive as a competitor's headline announcement, even when it creates more value. And your judgement about the right thing to do for the organisation can be distorted by what you think you should be seen to doing.

A: Yes. People anticipate how they might feel about a decision in the future, and change their choices now to avoid that regret. The executive is asking, consciously or otherwise: what if everyone else moves and I'm the person who failed to act? The perceived risk has flipped. A few years ago, an unsuccessful AI investment could make a leader look reckless. Now, failing to engage can look like the greater leadership failure, so doing something feels psychologically safer than doing nothing, even when conviction about the right action is low.

A: Three things. The first is forgetting your purpose, what you're trying to achieve. The second is the opportunity cost of money and attention diverted from better uses, which takes agency away from your people. And the third is what it does to capable teams. If you've hired good people, and you're not listening to their view of what's possible, you either fracture the team or make people disengaged, and that’s very hard to reverse, especially when they’re an expert in their field and you’re not.

A: Agency and conviction: knowing which capabilities matter to the organisation, what you want AI to change, and what you deliberately choose not to automate. Then the discipline to ask whether a new development changes that conviction, accelerates it or redirects it. Good leaders ask the questions others assume will slow them down, and answering those questions is what speeds things up. FOMO works against that because it pushes you to act before you've asked them.

A: Replace the question “are we behind on AI?” with a better one: behind whom, on what specific capability, and according to what evidence? Then ask why that capability matters to your economics, your customers or your mission, what your experiments are designed to teach you, and what evidence would make you scale or stop.

A: It starts with being deliberate about what you offload. We've always externalised cognition, to writing, calculators and search engines, but AI lets us offload synthesis, interpretation and elements of judgement itself. So the strategic question is which cognitive capabilities an organisation can afford to stop practising. It also means keeping the capacity to interrogate what the system gives you. If reasoning is delegated without that, you gain efficiency while gradually losing the judgement you need to recognise when the system is wrong.

There's something about how the human cognitive system works that's quite valuable, and we shouldn't lose it. Your people have expertise for a reason. Ultimately, this is human and machine cognition working together. And some anxiety is rational during a genuine technological shift; the aim is to convert it into better questions rather than indiscriminate activity for the sake of it.

Rakhi Rajani is a chartered psychologist, computer scientist and advisor to Nearform. Want to pressure-test your own approach to AI, and make sure it's strategy, not fear, driving it? Talk to us today.