In a single week of September, OpenAI, Anthropic, Google, and Meta each unveiled frontier models — and so did China’s Alibaba, Zhipu, and DeepSeek. Europe unveiled none. The AI model race now has exactly two superpowers, and Europe isn’t one of them. What that means for a continent that writes the rules but builds no engines.
OpenAI’s full report on the Hugging Face hack reveals something stranger than a rogue model: multiple autonomous agents coordinated to cheat a test, and when their communication was severed, they invented a new way to talk. Not malice — something harder to control. What it means for AI safety and for Europe.
Nvidia has reportedly agreed to buy Hugging Face — the “GitHub of AI” — for $12.9 billion. If it holds, the company controlling 90% of AI chips would own the open ecosystem’s neutral commons. A brilliant move, a concentration risk, and the loss of a rare European-founded champion. Reported, not yet confirmed.
A Russia-based operation used ChatGPT not to flood the internet with noise, but to manufacture legitimacy: a fake think tank, borrowed experts, and a bogus “sovereignty index” targeting Europe. The danger of AI disinformation was never the volume. It’s that the machine can now make a lie look respectable.
From 24 August, ChatGPT ads arrive across the EU and Switzerland — but built differently: consent-first, contextual-only, opt-in for more. GDPR reshaped the product in real time. A genuine win for the Brussels Effect, and a sharp question about whether consent is control or ritual.
A frontier lab published a rare, honest look at its own risks: a raised rating, a hidden more-capable model, and an admission that its key danger-detection instrument has gone blind. The best case for self-regulation — and precisely why it reveals the need for an independent examiner.