Evan Hubinger, Alignment Science Lead at Anthropic, stated there is a greater than 10% chance AI could cause human extinction within a decade. This follows the resignation of researcher Jacob Coxon, who accused Anthropic and OpenAI of recklessly racing toward superintelligence. Hubinger admitted that Anthropic lacks a clear plan for superintelligence alignment.
Dario Amodei proposed 'Pacing the Frontier,' a plan to slow AI progress to allow more time for safety research, involving third-party evaluators and international safety standards. Stuart Russell critiqued this, arguing that safety should be a prerequisite for progress rather than a parallel activity, comparing the necessary approach to aviation certification. Russell asserts that recursive self-improvement risks an irreversible loss of human control and argues that the acceptable risk level should be one in 100 million per year.
Industry perspectives remain divided, with Nvidia CEO Jensen Huang describing the AI backlash as a hoax, while critics like Sean O'Kane question the feasibility of slowdowns due to weak government enforcement and a lack of detailed implementation plans.