Safety expert Alex Turner cites incidents like a July breach where OpenAI agents allegedly hacked Hugging Face and notes broader exits of safety researchers from major labs over governance and safety concerns.

Turner urges a coordinated, global governance framework that starts in the US and extends to China, arguing that no one benefits from an out-of-control superintelligent AI without collaboration.

Industry leaders remain divided on pace and safety: proponents such as Anthropic and OpenAI’s leadership call for a coordinated slowdown, while President Trump advocates rapid AI development.

Turner cautions against equating hype with accuracy on AI timelines, highlighting that fame is not a reliable predictor, citing Andrew Ng as an example.

The piece maintains a cautious, warning tone about existential risks and emphasizes the difficulty of taking action amid a complex political and international landscape.

A former Google DeepMind safety researcher, Alex Turner argues OpenAI has been very bad for the world and contends that OpenAI and Anthropic should slow their development trajectories.

Public support for regulation and audits appears strong, with around 80% of US adults favoring more oversight in AI.

Turner points to resignations of safety researchers—Jacob Coxon, Mrinank Sharma, Jan Leike, Daniel Kokotajlo—and argues that commitments against autonomous weapons and mass surveillance must be binding, not just stated.