Mon, 2 Nov 2026
Home › AI
Anthropic · Claude

Anthropic revises Claude's "Constitution" and hints at AI consciousness

The new constitution runs to 80 pages and sets four core values. It ends by saying: "Claude's moral status is deeply uncertain."

Osamah Ismail · Founder & Editor-in-Chief
· 1 min read · Updated:
Anthropic revises Claude's "Constitution" and hints at AI consciousness
Image: Anthropic

On January 21, Anthropic released a new version of Claude's "Constitution", timed with Dario Amodei's appearance at Davos.

What's new?

  • The document runs to 80 pages
  • It keeps most of the 2023 principles and adds more nuance on ethics and user safety

Four core values

  1. Broadly safe
  2. Broadly ethical
  3. Compliant with Anthropic's guidelines
  4. Genuinely helpful

Safety

The constitution says Claude should "always refer users to relevant emergency services or provide basic safety information in situations that involve a risk to human life". Discussion of developing bioweapons is strictly prohibited.

Consciousness?

The document ends with this line: "Claude's moral status is deeply uncertain."

Sources: Anthropic, TechCrunch
Osamah Ismail

Founder and editor-in-chief of Kurd Developers. Phone and tech news on YouTube, Instagram and TikTok.