AI at Work: Efficiency Revolution or New Mental Health Challenge?

Generative artificial intelligence (AI) is now one of the most powerful drivers transforming our working environments. While its promises of productivity and automation are widely documented, its impact on workers’ psychological balance has so far remained largely unexplored.

To shed light on these issues, teale, a leading workplace mental health prevention platform, joined forces with Dr Justine Massu, a researcher in organisational psychology, and the innovation studio Tomorrow Theory. Together, they launched this first barometer dedicated to the links between AI and mental health, giving organisations practical guidance for reconciling technological innovation with psychological well-being.

[hook1]

The barometer is based on a rigorous survey conducted between July and September 2025 among 1,956 respondents working in France, representing the active French population.

The study notably uses the WHO-5, a standardised indicator developed by the World Health Organization, to assess employees’ overall psychological well-being over a given period. The objective is clear: to understand whether generative AI acts as a driver of fulfilment or a source of vulnerability.

Generative AI Adoption in Companies Is Moving at Different Speeds

The barometer’s first finding challenges the idea of the massive and immediate adoption often portrayed in the media. Although generative AI is highly visible in public debate, its practical use remains uneven.

• Still mainly occasional use: Only 38.1% of working people can be considered regular users, using AI weekly or daily. Most respondents use it occasionally.

[hook2]

• Generative AI as a social marker: Adoption is not socially neutral. The most regular users are, on average, younger, more highly educated and more likely to work in intellectual or scientific professions. By contrast, people in manual roles or direct services to individuals report much less frequent use.

• Tasks focused on cognitive support: Generative AI is mainly used where the return on effort is immediate: writing, information research, rephrasing and structuring. More complex uses, such as programming or data analysis, remain limited to a minority.

Generative AI Is Reshaping the Relationship with Work: Productivity, Autonomy and Usefulness

Generative AI does not simply perform tasks. It profoundly changes how employees perceive their work. The study highlights a central “psychological triad”:

• The productivity lever: AI is perceived as an enabler that reduces the cognitive load associated with repetitive tasks. It frees mental space for more meaningful activities.

• Greater autonomy and sense of competence: By supporting execution, AI meets the psychological need for competence and autonomy. Users feel more capable and better equipped to handle their responsibilities.

• The occasional-use paradox: Negative perceptions, including loss of meaning and a sense of uselessness, are concentrated among those who use AI least. For intensive users, AI is a stabilising resource; for occasional users, it can feel like a threat to their professional identity.

Generative AI’s Impact on Users’ Mental Health: Reassuring Findings

The heart of the study concerns psychological well-being. Contrary to fears about technostress or dehumanisation, the short-term figures are broadly positive.

• Satisfactory well-being scores: Generative AI users record WHO-5 scores between 60 and 67 out of 100, which the WHO considers satisfactory.

• A positive correlation with frequency: Strikingly, the more frequently generative AI is used, the higher the well-being score, rising from 62 for monthly use to 67 for daily use. This suggests that mastering the tool acts as a protective factor rather than a source of anxiety.

• Different impacts depending on the task: Delegating project management is particularly associated with greater vitality at work. Using AI for data analysis or translation is also correlated with better overall well-being.

Latent and Invisible Risks

Although the current data are encouraging, the barometer’s authors warn of latent effects of generative AI use at work that may only appear over time.

• The risk of deskilling: Partial automation can gradually erode skills if users stop practising actively. Some respondents already express concerns about the long-term negative impact of generative AI on their sense of competence.

• Relational substitution: If generative AI helps formulate feedback or prepare meetings, it must not replace authentic human interaction. Excessive reliance on AI for social relationships could create an “illusion of well-being” that conceals an impoverished collective.

• Work intensification: The time saved through generative AI may be immediately absorbed by higher expectations for volume or speed, cancelling out the mental health benefit.

Conclusion: Generative AI as a Company Culture Amplifier

The final conclusion of the barometer by teale, Tomorrow Theory and Dr Justine Massu is that generative AI is not inherently a source of well-being or ill-being. It acts as an amplifier. It strengthens positive dynamics where use is well understood and supported, but weakens organisations where it creates loss of meaning or identity insecurity.

Three action levers emerge for companies:

1. Reduce the technology divide: Encourage equitable access and make AI use visible and supported for all profiles.
2. ‍Move from passive to critical use: Train employees not only to use the tool, but also to reflect actively on their own expertise.
3. ‍Use human feedback to guide adoption: Do not focus solely on productivity metrics. Track qualitative indicators such as mental fatigue and sense of competence to adjust AI integration.

Generative AI can be a genuine lever for fulfilment if it is treated as support, not as a substitute for human intelligence and sensitivity. This is the condition for technological transformation to become a sustainable mental health lever.

Discover our five strategic levers for turning AI into an ally of your company culture in our second article about the barometer published by teale, Tomorrow Theory and Dr Justine Massu.