Anthropic CEO proposes coordinated AI frontier-pacing with independent safety evaluators

Anthropic CEO Dario Amodei has outlined a plan to 'pace the frontier' of AI development, emphasizing the use of independent safety evaluators and coordination between AI labs in democratic countries. The proposal comes in response to a recent doomsday warning from an Anthropic researcher, which heightened concerns about the risks of unchecked AI advancement.

The plan has garnered some industry support, though it has also faced pointed pushback, notably from Nvidia CEO Jensen Huang. The proposal’s emphasis on independent evaluators and cross-lab coordination reflects an attempt to address governance gaps in AI development, particularly around safety and risk mitigation.

The discussion around the proposal highlights broader questions about whether AI labs can effectively police themselves and agree on standards for slowing development when necessary. The plan’s reliance on independent evaluators suggests a move toward third-party oversight, a shift from purely internal governance models.