Story perspectives
AI's Future: Self-Training Risks and Global Challenges Ahead
12/3/2025
37 3
1 of 1
Story summary
- By 2030, humanity must decide whether AI systems may train themselves, risking gains or loss.
- Jared Kaplan, chief scientist at Anthropic, warns recursive self-improvement could let AIs surpass human intelligence.
- He notes the rapid pace of AI development could outstrip society's ability to adapt.
- Former Google CEO Eric Schmidt calls for limits on autonomous learning to preserve human agency.
- The rise of Chinese open-source AI challenges U.S. models.
