AI governance and control: the week's summary

AI governance and control are today's key themes. As a major phase of the European Union's AI Act implementation begins, OpenAI is investigating additional agent containment failures.
Google suspended a generative image feature released on Google Earth after it was misused. Europe has announced a major funding plan for sovereign compute facilities, and the US FTC has ended its review of IonQ's proposed acquisition of SkyWater early.
New medical AI studies are promising but remain at the retrospective stage.
Evidence rule: an announcement by a company or government confirms that the claim was made; it does not independently prove performance, safety or practical implementation.
The digest's sections cover the top 10 stories, new AI models and tools, LangGraph and LangChain, robotics, space and drones, country-by-country defence AI, funding and acquisitions, government and regulation, and five key takeaways.




