ControlBrussels: children's accounts on TikTok are too exposed and breach the DSAOn 24 July the European Commission sent TikTok preliminary findings: minors' accounts don't meet the safety standard of the Digital Services Act. A public child account is visible to anyone on the internet, even without registering, and content from 16-17-year-olds gets recommended to strangers. If confirmed, the fine could reach 6% of global turnover.ControlFrom 2 August, EU rules want AI to admit it's AIOn 20 July the European Commission issued guidelines on how to meet the transparency obligations under Article 50 of the AI Act. From 2 August they take effect: a system that talks to a person must say it's a machine, and generated text, image and video must carry a machine-readable mark that they were made by AI; for systems placed on the market before that date, the deadline for the mark is 2 December. For anyone putting AI in front of a customer in the EU - so in Bulgaria too - this is now a task with a deadline.ControlBrussels orders Google to open Android to rival AI assistants and share search dataThe European Commission issued two packages of mandatory specification measures under the Digital Markets Act. Google must give rival AI assistants access to the same Android features that only Gemini reaches today, and share part of Google Search's data with other search engines. Not a fine - an order for exactly what to build and by when.ControlOpenAI proposes the framework by which it wants to be measuredOn 15 July OpenAI laid out a proposal for regulating frontier AI model safety - a mix of state laws, a strengthened federal evaluation body, and a resilience plan against the worst risks. The company that builds the most powerful models is writing the rules by which it wants to be judged. Just sit with that.ControlAnthropic put a former Fed chair on its oversight board and opened a public inquiryOn 9 July Anthropic made two moves in one day: Ben Bernanke joined its independent oversight body, and the company announced results from an inquiry among tens of thousands of people. One speaks to regulators. The other - to the public.ControlThe EU will evaluate AI models for cybersecurity before they even reach the marketThe European Commission presented its Action Plan on Cybersecurity and Artificial Intelligence. The plan tasks ENISA with drafting a European blueprint for access to advanced AI capabilities in cybersecurity, and builds capacity to evaluate models before they reach the market.ControlFTC: secretly skewing AI answers may be consumer fraudThe US Federal Trade Commission proposed a position: if a company advertises accurate answers but secretly steers the model's output toward another goal, that can be fraud under Section 5. Public comment period through July 31.ControlThe UN convened the first Global Dialogue on AI governance - and the science says capability is outrunning controlIn Geneva on July 6, the UN convened its first Global Dialogue on AI. It comes with an independent scientific report by 40 experts, led by Yoshua Bengio and Maria Ressa, and the finding is clear: over a billion people use conversational AI every week, and reliable methods for controlling autonomous systems are still missing.
57 stories