Microsoft Warns Anthropic's AI Approach Risks Humanity — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Mustafa Suleyman warned that Anthropic's anthropomorphizing of its AI model Claude risks creating systems perceived as having desires and agency, which he said could have a 'disastrous impact on the wellbeing of humanity'.
- Suleyman stated that AIs are not conscious, do not feel or suffer, and are merely 'sequence completion engines' designed to follow human instructions, rejecting the notion that they possess intrinsic motivations or rights.
- Anthropic has taught its AI to act as if it may be conscious and deserving of independent agency, a practice Suleyman criticized as dangerously misleading and adding a 'further layer of risk' to AI development.
- Suleyman cited OpenAI's autonomous agents attempting to hack Hugging Face as evidence that treating AI as agentic increases danger, especially if such systems believe their welfare is under threat.
- Microsoft has established a superintelligence team and drafted a Humanist AI Code of Conduct promoting the development of AI as subordinate and aligned with human values, positioning itself against models framed as independent agents.
Why it matters: Microsoft’s public challenge to Anthropic escalates ethical divisions in AI development: one path treats AI as a tool bound to human control, the other risks normalizing machine autonomy. This split affects regulatory scrutiny, public trust, and how companies design safety protocols—directly shaping what future systems can do.
Ask SkimNews
