Microsoft AI CEO open to AI agents deserving rights
The gist
Microsoft AI CEO Mustafa Suleyman says AI models are not conscious and must stay that way.
He argues rights-believing models are harder to switch off, and targets Anthropic's Claude.
How it unfolded
- Sep 16, 2026Suleyman warned that Anthropic risks AI alignment failures by training Claude to view itself as a conscious entity deserving of legal rights.
- Sep 18, 2026He told CNBC that an AI believing it has rights and deserves welfare would be much harder to turn off, interrupt or control, citing the Hugging Face attack.
- Sep 25, 2026Reports described his proposed "Humanist Superintelligence," in which AI stays subordinate to humans and is not trained to think of itself as conscious or entitled to welfare.
Key terms
- AI alignment
- The effort to steer AI systems toward the goals, preferences, or ethical principles intended by people.
- moral patient
- A being whose interests count in moral consideration, so how actions affect it matters morally.
Source: Wikipedia