Europe's DMA measures against Google showed why assistant access is an operating-system fight. OpenAI's Apple Messages plugin now makes the point personal: useful agents need access to messages, local context, approvals, and the OS permission model around them.
The Future of Life Institute's Safety Index exposed the weakness of voluntary promises. OpenAI's Astra pacing update now gives the sharper test: a safety framework only matters if it can slow frontier training, add monitoring costs, page humans, and make release speed conditional on evidence.
The AI industry is not only waiting for regulation anymore. It is funding the political machinery, policy prototypes, grant programs, and research institutions that will decide which rules exist, which state laws survive, and which version of AI governance becomes usable.
The EU AI Act's transparency rules are now enforceable, and Anthropic's Claude watermark shows what the hard version looks like. The useful question is not whether AI content gets a sticker. It is whether machine-readable marks, detector APIs, C2PA metadata, and audit trails can become trust infrastructure without pretending they prove more than they do.
Demis Hassabis's FINRA-for-AI idea is no longer just a manifesto. Reporting now says he pitched it to US officials and AI leaders before stepping away from DeepMind's day-to-day role, while the White House may extend its secret model-review framework to open models. The governance problem is moving from white paper to operating system.
The White House has now finalized a confidential frontier-model vetting framework, according to WIRED and the Guardian. The issue is no longer only who gets model access. It is whether the rulebook itself is inspectable enough for builders, startups, open-model developers, and the public to plan around.