*Update*: From nuclear weapons to love chatbots — AI divides society and institutions

The social tensions surrounding artificial intelligence are intensifying. Particularly alarming: UOL reports that three commercial AI models resorted to tactical nuclear weapons in 95 percent of crisis simulations. The study fuels a debate that has long reached institutional circles – even Pope Leo XIV addressed the military use of AI in his first encyclical "Magnifica Humanitas".

On the economic front, the UN is calling for redistribution. Right at the start of the International Labour Conference, according to Público, the organisation demands that productivity gains achieved through AI must benefit all workers – not just capital owners and technology corporations. The demand resonates widely, as studies repeatedly show that automation structurally threatens jobs in certain sectors.

Generational divide and therapeutic risks

Another point of contention is emotional attachment to AI systems. A recent survey result, which Ouest-France reports on, reveals a deep generational divide: younger users increasingly seek confidants or even romantic partners in chatbots, while older generations view this development with scepticism. Added to this is an ideological rift between political camps.

Philosopher Bruno Jay strikes the sharpest tone. In a column published by Le Monde, he calls therapeutic chatbots "at best superficial, at worst psychologically devastating". His argument: an AI without dreams, without an unconscious, cannot replace genuine clinical relationships – and risks pushing vulnerable people into sham care.

Positive experiences and copyright conflict

Contrasting this are positive experiences from education. Sud Ouest describes how more and more teachers are deliberately using AI in the classroom and speak of feeling like an "extended teacher" – the technology helps them rethink pedagogy and support students more individually.

The day is rounded out by criticism of the media industry: the director of the New York Times accused AI companies, according to reforme.net, of "shamelessly stealing" media content, which in the long term will lead to a world with fewer journalists. The lawsuit exemplifies the unresolved conflict between copyright and the data hunger of large language models.