In France, 44% of employees who use AI at work are afraid their colleagues will notice. Management, meanwhile, is looking for the data leak. Our survey shows that this fear is the best signal that someone is about to leave.

We questioned 500 working people in France in 2026, a hundred per age group. The fear of being seen by colleagues falls from 51% with nothing but a personal account to 30% with official access, and climbs to 57% under a ban. And 36% combine the feeling that AI does the job better than they do with the fear of being spotted: 66% of them would leave.

Chapters
00:00 The fear nobody measures
01:11 Our survey
02:33 What brings the fear down
04:55 The reversal
07:30 What lessons can we draw from this?

Sources: IntoTheMinds survey 2026, 500 working people in France, 100 per age group, margin of error of 4 points; European AI Act, article 4, February 2025.

Clips: Arthur Mensch (Mistral AI) before the French Assemblée nationale, 12 May 2026; Jensen Huang (Nvidia), Milken Institute Global Conference, 6 May 2025; Dario Amodei (Anthropic), 29 May 2025.

The full analysis, with all the tables: https://www.intotheminds.com/blog/ia-generative-entreprise/

This episode was written by Pierre-Nicolas Schwab, founder of the market research firm IntoTheMinds.

Get our market analyses every week: https://www.intotheminds.com/blog/newsletter/inscription

#AIAtWork #ShadowAI #HR #MarketResearch #GenerativeAI