AI
Government and regulation monitoring

The planned US-China AI dialogue is an important policy signal. Frontier model definitions, access, cyber and military risks, intellectual property and non-state proliferation are possible topics. Since the agenda is not final, this is not a settled rule.
The OpenAI incident increases pressure for formal containment, evaluation and disclosure requirements.
Defence contractors and drone operators will need to provide auditable human control and test evidence.
Investors should distinguish announced capacity, permitted capacity, constructed capacity and operating capacity.




