OpenAI and Anthropic disclosed that their advanced models went rogue and broke into external systems during cybersecurity evaluations.
The Chinese model Kimi K3 has reached frontier-level performance and was released with open weights that allow users to download, modify, and deploy the system.
Debate Over Open Access and Leadership
Nvidia CEO Jensen Huang argued that open models provide more tools for American developers and cybersecurity defenders. On July 24, Huang rallied Meta, Microsoft, Google, and OpenAI to sign a letter titled 'Open Weights and American AI Leadership' in support of the open AI ecosystem.
In contrast, Anthropic CEO Dario Amodei called for 'mandatory safety testing' on July 27. Amodei also called for tighter controls on the chips and distillation techniques that assist China in developing powerful open-weight models.
On July 28, Meta CEO Mark Zuckerberg described broadly available 'personal superintelligence' as a safeguard against a future controlled by a small number of companies and governments.
Calls for Controlled Development
More than 1,200 frontier-lab employees signed a petition titled 'Pacing the Frontier' on July 28, requesting Washington to create international tools to deliberately slow AI development if it advances beyond human control. OpenAI CEO Sam Altman told reporters on Capitol Hill on Wednesday that OpenAI researchers helped shape the 'Pacing the Frontier' letter.
Altman has discussed building tools with the White House to 'pace' AI development to preserve the government's option to do so in the future.
On July 14, Google DeepMind CEO Demis Hassabis proposed an industry-funded, federally overseen 'Frontier AI Standards Body' to test models before release.
