AI Safety

Latest news and articles about AI Safety from NEX English

Warning Issued as AI Approaches Self-Improvement: Humans Risk Losing Control Soon

Anthropic co-founder Jack Clark has warned that artificial intelligence is approaching "recursive self-improvement," where models could evolve without human intervention. The company revealed that <strong>80%</strong> of its programming work is currently handled by the Claude system, urging global safeguards before autonomous development becomes unstoppable.