Floridi urges critical, responsible adoption of AI, not blanket bans

Floridi’s interview matters because it sketches a middle path between “pause AI” maximalism and uncritical acceleration. He argues that AI should be seen as a powerful informational agency that remains non-conscious, and that the core ethical move is to keep humans accountable for its use. That includes resisting ideas like granting full legal personhood to AI-run corporations, which would make it harder to assign responsibility when things go wrong.

He also pushes on a subtler risk: environments being redesigned for machines rather than people. From SEO-optimized headlines to physical spaces shaped around robots, Floridi warns that we are already adapting our world to AI constraints instead of the other way around. For an AGI-focused audience, this is a reminder that path-dependent choices in UX, law and infrastructure will shape how controllable future systems are long before we cross any formal AGI threshold.

His call for an “ethics of the negative” in international agreements, focused on what AI systems must not be allowed to do, is tactically important. It is often easier to coordinate globally on red lines than on positive visions of progress. If that framing gains traction, frontier labs may find themselves operating inside a clearer but narrower envelope of acceptable behaviors and use cases, even as they keep pushing capabilities.