The Cognition Brief: technology, humans, and what changes

Sonya Cullington:

Compliance Without Conviction

If you work in communications, you have probably led or supported an AI adoption programme that hit all its adoption targets. Everyone attended the functional training lunch and learn sessions. The launch went well, the content landed, the usage data looks credible. And yet you may also have a nagging sense that something did not quite take hold. That the change you were asked to deliver happened on the surface but not underneath it.

Most change models assume that trust follows from good engagement. Explain the case clearly, involve people early, listen properly, respond honestly, and the organisation gives people a fair chance of coming with it. Much of the time, that is true. With GenAI, it is not enough.

Why GenAI is a different kind of change

When a new system lands, people are rarely only asking whether it works. They are asking, often to themselves, whether it is safe to rely on, whether it will make them look less competent and whether the organisation will stand behind them when something goes wrong.

With most change programmes, that assessment sits in the background and often settles as people get used to the new way of working. GenAI is different. People do not respond to it like a new platform or a redesigned process. They start to judge it, and they start to judge the organisation that introduced it. Does leadership understand the pressure this creates? Will they be honest about what it cannot do? Will they notice if something goes wrong?

That is why trust in AI adoption is not built by information alone. A good internal campaign can explain what the tool does, why it was chosen, where support sits, and what safeguards are in place. All of that is important of course, but people still need to feel, through experience, that the organisation understands what this technology asks of them, and will be honest when things become difficult.

The pattern 

This is the pattern many communicators recognise. The dashboards are not alarming. The launch content has gone out and leaders have said the right things. People are using the tool often enough for the programme to look credible. And yet, in the places where judgement, confidence and trust matter most, the change has not really taken root.

You see it in the small behaviours. A team uses the AI tool when the programme team is watching, then quietly goes back to the old way of working. A manager mentions AI outputs in meetings but makes the real decision elsewhere. An organisation reports healthy adoption while its people build workarounds around the edges.

These are not signs that the communication has failed. They are signs that trust between the organisation and its people has not yet been built where it matters. People may have understood the case for adoption. They may even agree with it but they have not yet reached the point where they believe the organisation genuinely understands what it has introduced, or that it will support them through the difficult parts of the adjustment.

Communications professionals often know when this is happening. They may describe it as a cultural issue: we have not brought people with us, there is a mindset shift needed, we need to make experimentation feel safe. That language is not wrong. It is often pointing to the right problem. But it does not explain why the problem is happening.

The questions that are not being asked

Most communications briefs for AI adoption start with the technology. What does it do? How will it change workflows? What are the benefits? What concerns might people have?

The questions that tend not to appear are the ones that determine whether any of that communication will land.

The questions that matter most are often the ones that feel harder to ask. How much trust already exists between leaders and staff, and is it strong enough to carry this change? Where might people feel their judgement, autonomy, or professional standing is being threatened? If someone said, openly, that they did not trust the AI, would that be treated as useful intelligence or as resistance?

These questions sit in the space between communication and psychology. They are not about messaging. They are about the conditions that need to exist before messaging can work.

A starting Point, not a framework

Strategic communications have earned its place in conversations about organisational change. The next challenge is to notice where the familiar playbook stops being enough. AI adoption is one of those places. The usual tools can get the launch right, the narrative right, and the engagement plan right, while still missing what is happening underneath.

The missing layer is psychological. It is about how trust between organisations and their people holds or fractures during technological change, how they respond when technology starts to behave like a social presence, and what happens to professional identity when tools begin to shape the thinking, not simply support the work.

This column will explore that layer. Each piece will start with a problem communications professionals already recognise, then look beneath the surface at what is happening psychologically. The aim is to help the profession ask better questions earlier, so the next AI adoption programme is not just better communicated, but more honestly understood.

+++

Sonya Cullington is a cyberpsychologist and digital policy advisor.  This is the first of a series exploring the psychological layer of technology adoption and its consequences for communications practice.