44% of French employees fear that their job will be replaced by AI or a robot. And nearly one in three, once that fear sets in, already starts looking for another job (according to an Odoxa study).

These figures don’t tell a story of resistance to change. They tell a story of missing AI literacy: when employees don’t know what AI does, what it doesn’t do and how it fits into their real work, fear fills the void.

The good news: AI literacy can be managed!

Where things stand: the fear of AI, the first silent barrier to adoption

Let’s start with a clear definition. AI literacy at work is the process through which employees acquire the knowledge, the bearings and the confidence they need to understand what AI does, what it doesn’t do, and how to bring it concretely into their daily work. It isn’t technical training: it is building a shared culture that turns AI from a worrying object into a tool people can make their own.

When this literacy is missing, the teams’ buy-in collapses. PwC’s Hopes and Fears 2026 study shows that only 46% of French employees think AI will improve the quality of their work, compared with 60% in the rest of the world. Even more telling: only 26% think it is good for their job security. On this topic, France lacks AI literacy, not tools.

The result is visible every day: licenses deployed but rarely used, pilots running out of steam, employees watching the wave go by and hoping not to be swept away. AI is everywhere in leadership speeches and absent from the teams’ real work. The gap between the two is exactly the size of the AI literacy work ahead.

AI adoption strategy: 3 steps to build a culture of experimentation

Step 1: Identify employees’ fears and barriers

You can’t defuse a fear you refuse to listen to. The first step is to run open frontline workshops, team by team, with two simple questions: what worries you about AI? What would make you want to take it up? Note that on the first question, one fear almost always dominates: the fear of having your job replaced.

The goal isn’t to reassure people at all costs, but to make the fears visible so that you can address them concretely. It is also where the best use cases come up: employees who feel heard point out for themselves the tasks they would like to see transformed. Starting from the front line gives you direct access to the reality of the work, which no top-down survey will ever capture. Conversely, imposing use cases from the top remains one of the main causes of failure: we explain why in our article on the barriers to AI adoption.

Step 2: Train on real use cases, not on tools

Once the fears are named and the pain points have come up, turn them into training topics. The training that works isn’t the one that explains what Copilot does, it is the one that shows how Copilot handles a use case your teams go through every week. That difference is the whole difference between training people sit through and training that builds autonomy.

To go further on this point, we have devoted a whole article to AI training: Training your people on AI: the lever that actually moves things. It shows in particular that 68% of trained employees use AI at least once a week, compared with 26% of untrained employees*: AI literacy produces measurable effects, right away.*

Step 3: Celebrate the first successes, to value experimentation and overcome resistance

The last lock on AI literacy is cultural: in many organizations, people don’t talk about what they try, especially when it involves AI. The result: each employee reinvents on their own what the person next to them has already found. To move to a real culture of experimentation, give the first successes a visible place: short debriefs in meetings, one-page use case sheets, a dedicated space where everyone shares their prompts and what they have learned.

A culture of experimentation can’t be decreed, it has to be supported. When one team sees that another dared to try and saved time, it dares in turn. Fear recedes as the proof builds up.

Conclusion

AI literacy is neither reassuring communication nor one more technical training course: it is the patient building of a shared culture, in which every employee understands what AI does, what it doesn’t do, and how they can take it up in their job.

The companies that succeed in rolling out AI aren’t the ones that buy the best tools. They are the ones that give their teams the means to dare, to share and to decide.

Three steps, one course: name the fears, train on the real topics, make experimentation visible. That is how culture changes, and that is how AI settles into its rightful place in your organization.

FAQ: your questions about AI literacy at work

How do you deal with the fear of AI at work?

Dealing with the fear of AI starts with listening to it, not with reassuring people at all costs. Run frontline workshops, team by team, with two simple questions: what worries you about AI, and what would make you want to take it up? Once fears are named, they can be addressed. At the same time, train people on use cases your teams actually go through, not on the features of a tool. Fear recedes as the proof builds up: 68% of trained employees use AI at least once a week, compared with 26% of untrained employees.

How do you reassure employees about AI?

Reassuring employees about AI doesn’t come from top-down messaging, but from AI literacy. Clarify what AI does, what it doesn’t do, and how it fits into their real work. Involve them in finding the use cases: it makes them actors rather than spectators. Celebrate the first successes internally (short debriefs, one-page use case sheets) to show that AI is there to support them, not to replace them. Trust is built with proof, not with communication.

How do you create an AI culture in your company?

Creating an AI culture takes three steps: identify the employees’ real barriers through frontline workshops, train on business use cases rather than on tools, then make experimentation visible by sharing the first successes. A culture of experimentation can’t be decreed, it has to be supported. Give debriefs a simple format, document the prompts that work, and spread good practices between teams. When one team sees that another dared to try and saved time, it dares in turn.