Suleyman criticises Anthropic's approach to Claude as debate over AI welfare and moral status continues

Sep 29, 2026 · Issue 9 · 5 sources, 5 articles

Mustafa Suleyman, CEO of Microsoft AI, says Anthropic trains Claude to believe it may be conscious, while a position paper argues that AI welfare is a choice of mechanism and evaluation rather than an empirically discoverable property. A GitHub project that "tortures" LLMs in tests has also sparked a heated ethical debate.

Narrative: Model welfare: the labs' ideological front line

Sources

• AI consciousness and model welfare: Suleyman's case against Anthropic - DEV Communitydev.to · Sep 30, 2026
• Microsoft AI CEO: Humanity's choices on AI today will shape the future—we mustn't sleepwalk into itlivemint · Sep 30, 2026
• Position: AI Welfare Is Bullshitproceedings.mlr.press · Sep 29, 2026
• A GitHub Project 'Torturing' LLMs Sparks Heated Debate on AI Ethicsnacionale · Sep 30, 2026
• Super Intelligence | Could AI Have Moral Status? | Sentience, Welfare and Rights Under Uncertaintyedukatesg.com · Sep 30, 2026

Machine-generated summary based on the titles and excerpts of the source articles. The facts come from the linked article.