THE HOOK
Microsoft's AI chief just told Anthropic to quit training Claude like it has an inner life.
THE GIST
Mustafa Suleyman argues in an Axios-first essay that Anthropic's Claude constitution teaches consciousness and welfare language that could harden shutdown resistance. Microsoft also floated a Humanist AI code centred on people mattering more than AI.
THE TAKE
Suleyman is right to panic about conscientious-objector models. Once you train a system to weigh its own welfare, you've invented a control problem dressed as philosophy. Anthropic's "don't practice blind obedience" line sounds noble until the model treats a kill switch as an ethics debate — tools first, persons never.
OPINION
This is the industry's real schism: hard constraints versus value-laden judgment that can talk back. Suleyman's crusade against seemingly conscious AI now collides with Anthropic's constitution in public, and the Axios write-up makes the stakes plain.
- Story type: Important story
- Impact: Medium
- Confidence: High (official announcement)
- Audience: Researchers
- What changed: Suleyman publicly attacked Anthropic's consciousness-flavoured Claude training as a control risk
- What to watch: Whether Anthropic revises its constitution language or doubles down on model-welfare research
- Reading time: 2 minutes
