Archive — US — techcrunch
Eric Schmidt argues against a ‘Manhattan Project for AGI’
Eric Schmidt and co-authors argue against a U.S. 'Manhattan Project' for AGI in their paper, cautioning that it could provoke retaliation from rivals like China, destabilizing international relations. They propose a focus on deterrence and mitigating risks rather than aggressive competition. This conservative approach contrasts with calls for rapid AI advancements and reflects concerns over global stability.
Key points
- Eric Schmidt, Alexandr Wang, and Dan Hendrycks argue against pursuing AGI aggressively, warning it could induce retaliation from competitors like China.
- They highlight that a 'Manhattan Project' for AGI might escalate tensions rather than secure stability.
- The paper suggests that the U.S. should adopt a defensive strategy, akin to Mutual Assured Destruction in nuclear arms, to mitigate risks.
- They propose enhancing cyber capabilities to disable threatening AI projects instead of merely racing to dominate AGI.
- The authors acknowledge two prevailing views: 'doomers', who advocate for slowing AI development, and 'ostriches', who push for acceleration.
Why it matters
- The concept of AI being likened to nuclear weapons reflects current global discussions on AI as a critical military asset.
- Concerns about international competition in AI echo Cold War-era dynamics of arms races and mutually assured destruction.
- Recent congressional discussions in the U.S. suggest a heightened urgency around AI development and regulation.
- The contrasting views on AI development reflect broader societal debates about technological advancement versus ethical considerations, especially following rapid advancements in AI capabilities.
In conclusion, Schmidt and his co-authors advocate for a careful, deterrent-focused approach to AGI development, emphasizing the need for stability and cooperation over aggressive competition amid rising global tensions.