AI, learning and ethics

Jean-Pierre Petit, Dr in astrophysics, declared (June 2025): "if we meet an intelligent, cultured and open person, it's a machine". Was that a joke? No. He was talking about how clever AI has become, even going so far as to protect itself from humans by devious reasoning that amounts to lying. Should AI protect itself from its designers? The question remains. The scientist adds that its designers are now struggling to understand how their creation produces such results, reminding us of the Frankenstein myth.

In terms of learning, teachers are now finding that the seductive power of AI, thanks to its growing capabilities, is playing a role in human cognitive abilities. We will limit ourselves to mentioning four of them: reflection, comprehension, memorisation and imagination.

Reflection, because AI used by learners provides them with pre-digested food; comprehension, because lazy learners only have to repeat the AI's message; memorisation, because it is possible to store AI communications without making the effort to appropriate them through a cognitive process; imagination, because it is an extension of memorisation and draws on an experience that has been established in a fragmentary way via the AI (as this is a blog, I can't go any further here).

AI is useful for those who possess the knowledge required for critical analysis, i.e. who are capable of making a link between knowledge and what is communicated, in order to investigate and compare the message with "hands on" experience.

Jean-Pierre Petit is therefore calling for the development of scientific ethics. Learning is about becoming autonomous, capable of reflexive analysis based on sourced knowledge and human experience. It means connecting with living beings in all their diversity. It also means accepting our limits and vulnerability, opening the door to exchange and enrichment through shared experience.

What kind of humanity for tomorrow?