by Thomas Lepeltier, Le Point, 20 September 2026
The political danger of AI would come not from a hostile machine, but from an assistant so helpful that we would end up entrusting everything to it. This is the thesis of ‘Guidance’, Antoine Buéno’s essay, which this article presents and discusses.
In brief
- In ‘Guidance’ (Tallandier), Antoine Buéno, a member of APALA’s expert committee, takes Huxley rather than Orwell as his compass: tomorrow’s power would rest on advice and incentive more than on fear.
- His equation: ‘big data + nudge + agent = spiritual direction’. The more competent and benevolent the AI agent, the more decisions we would delegate to it.
- The author welcomes a piece of foresight used as a thought experiment, without seeing it as a forecast to be validated point by point.
The article’s three arguments
A dependence born of the service rendered
Tomorrow, the AI agent will filter our messages, choose our information, and advise us on health or relationships. Each act of delegation seems harmless; it is their accumulation that could turn assistance into power.
Who will be speaking in our ear?
Big Tech, the State or citizens: depending on who controls the guidance, Buéno sketches several ‘AIcracies’, from authoritarianism to a ‘democrAIcy’. Technology alone does not dictate the political regime that will emerge from it.
Work, redistribution, biotechnologies
If AI made work scarcer, very different regimes might converge towards greater redistribution. Biotechnologies would then open up a choice between egalitarian use and an augmented elite.
‘The decisive question is then not whether AI will one day govern us, but how many decisions we will agree to delegate to it before realising that it has already begun to do so’Thomas Lepeltier, Le Point, 20 September 2026
Why this debate interests us
AI is at the heart of our work on progress, and two members of our expert committee are in dialogue here: Thomas Lepeltier presents Antoine Buéno’s essay. The book usefully shifts the debate: far from extinction scenarios, it raises a concrete democratic question, that of the decisions we delegate every day to useful tools, and of who steers them.
Questions for debate
- At what point does an assistant that advises us become a power that governs us?
- Who should control personal AI agents: companies, the state or users themselves?
- Can AI make economic planning effective, where it failed in the 20th century?
To go further on apala.fr
Summary written by Claude (model Opus 5.5), an artificial intelligence, for APALA, based on the article published in Le Point on 20 September 2026. The views expressed are those of the author. An error? Write to us.
0 comments