U.S. AI Strategy Needs to Shift From Speed to Societal Strength

The U.S. approach to AI -- move fast, win first -- is running into a wall. A commentary from the Carnegie Endowment for International Peace argues that mounting safety incidents and warnings from leading AI labs signal it is time to rethink what national AI advantage actually means.

Author Michael Mazarr points to two recent events as evidence: OpenAI paused the release of GPT-6.1 Astra over safety concerns, and Anthropic's IPO filing catalogued nearly 100 pages of risk while acknowledging that a catastrophic AI incident could exceed what the firm can insure against. These developments, Mazarr writes, expose the limits of a pure-speed strategy.

His alternative: replace the race to build more powerful AI with a campaign to build a stronger society for the AI era. Drawing on historical cases of lasting national advantage, he argues that durable power comes from resilient institutions, workforce readiness, and public trust, not from being first to deploy the most capable model.

Key Takeaways

• Speed-first AI strategy is on borrowed time. Safety incidents and expert warnings are increasing pressure on the U.S. to slow down and reconsider priorities.
• A societal strength model may outperform. Historical precedent suggests countries with strong institutions and adaptable populations sustain advantages longer than early movers alone.
• A deliberate pace creates opportunity. Slowing frontier model development could give the U.S. space to build governance, workforce capabilities, and public trust.

Read the full article on Carnegie Endowment for International Peace