AI safety legislation has a conflict-of-interest problem — Big US labs are warning of risk just as cheaper Chinese open-weight models erode their commercial edge

What happened
OpenAI has announced its support for a bipartisan effort in the US Congress to limit China's access to advanced AI chips and equipment. This development comes as Chinese companies release open-weight AI models that are less expensive and increasingly competitive. Around the same time, OpenAI introduced its latest model, GPT-6 Astra, which the company describes as its most capable and well-aligned system so far. OpenAI's internal tests indicate Astra performs better at following instructions and maintaining safety boundaries compared to previous models.
Key details
- OpenAI backs US legislation to restrict advanced AI chip exports to China.
- Chinese open-weight AI models are becoming cheaper and more competitive.
- OpenAI launched GPT-6 Astra in September.
- OpenAI claims Astra is its most capable and aligned model to date.
Why it matters
US lawmakers and technology companies face new decisions about how to balance national security, commercial competition, and access to AI technology. For US AI labs like OpenAI, the proposed restrictions could help maintain their market position as Chinese models become more affordable and widely available. Businesses and researchers relying on AI may see changes in the availability and cost of advanced AI models and hardware, depending on how export controls are implemented.
