The numbers worth knowing
Seven in ten employees in Germany expect AI use to lead to job cuts. More than a third are specifically worried about their own job. A quarter of respondents report having already lost their job because of AI – among under-25s, often in entry-level positions, that share rises to 43%. This concern is therefore not diffuse or overstated, but backed by the real experience of a substantial share of the workforce.
Where the concern outruns the evidence
At the same time: not every role is equally affected, and a blanket "AI replaces everyone" doesn't match the observed reality. Losses concentrate disproportionately on certain task types and entry-level positions, not evenly across the entire workforce. Naming that distinction honestly, instead of blanket-confirming or blanket-dismissing the fear, is the actual communication task.
The second, less-discussed concern: more error-proneness
Besides job worry, there's a second, less-discussed burden: more than half of people who use AI systems now feel their own work has become more error-prone or insecure – partly because accountability for AI-assisted decisions stays with the person who ultimately makes the call, while the AI itself bears no consequences. This concern is completely skipped over in many change communications, even though it's at least as widespread as job worry.
Practice section: participation instead of reassurance
The Fraunhofer Institute for Industrial Engineering recommends a participatory approach: employees should be involved in selecting and designing AI tools, instead of only being informed about decisions already made. That builds trust from the inside, not through outside assurances. Concretely, that means: ask before the rollout, don't just inform afterwards – and name the legitimate concerns openly (job risk in certain areas, error-proneness) instead of glossing over them with generic reassurance.