Who Controls the AI? Europe's Founders Are Asking Out Loud

Key points

• TechBBQ, an annual Nordic tech conference, wrapped up in Copenhagen on 27 August 2026 with AI sovereignty as its dominant theme.
• Anthropic's AI models Mythos and Fable became unavailable to users outside Europe earlier this year, disrupting software teams and sharpening the sovereignty debate.
• Signal president Meredith Whittaker warned that AI assistants built into phones and operating systems amount to a "data collection apparatus".
• Emad Mostaque, co-founder of Stability AI, said the person who controls the AI ultimately controls the country.
• Speakers framed the core question not as what AI can do, but what humans are willing to hand over to it.

The panels at TechBBQ this year kept arriving at the same uncomfortable place. Not "can AI do this?" but "should we let it, and who decides?"

The conference, held annually in Copenhagen for founders and investors across Europe, closed on 27 August. Its theme was "Emerging from Agency", a phrase that turned out to be well-timed. Earlier this year, Anthropic's AI models Mythos and Fable, two large language models (the technology behind chatbots like ChatGPT), were cut off for users outside Europe. The outage was a sharp reminder that much of the AI infrastructure European businesses depend on is owned and operated in the United States or China.

One startup executive told TechCrunch AI the disruption hit his software team hard. Another shrugged it off. Both agreed the underlying tension is real, even if it hasn't bitten hard yet.

What were people actually worried about?

The concern is sometimes called AI sovereignty: the idea that a country or region should control the AI models and computing infrastructure it relies on, rather than rent that power from abroad. Our 11 August story on Mistral's push for European AI ownership showed how at least one European lab is trying to make that practical.

Ellen de Brever, an angel investor and head of partnerships at the Novo Nordisk Foundation Cellerator, put it plainly at the conference. The debates, she said, "weren't about smarter machines, but about human judgment and who gets to stay in the driver's seat."

Meredith Whittaker, president of the encrypted-messaging app Signal, said the current AI wave is building a "data collection apparatus", and that AI companies use clever marketing to make users forget "about the collateral consequences" of handing over so much personal data. She pushed back specifically on AI assistants woven into operating systems, such as ChatGPT's integration with iMessage on iPhones, arguing it exposes far too much private data.

Her message to ordinary users was blunt: the convenience of an AI that reads your messages comes at a cost most people haven't been asked to price.

Emad Mostaque, co-founder of Stability AI, went further. "Sovereignty is the ability to resist power being exerted over you," he said. He warned that power over AI is concentrating in a handful of large labs, and predicted that "inevitably, every country will be run by AI" and that "the person that controls the AI controls the country." Stability AI raised $76 million in fresh funding just days before the conference, as we reported on 25 August.

What does this mean for ordinary people?

That is the part still being worked out. Mia Negru, advocacy director at the nonprofit Life With Artificials, wrote after the panel: "If intelligent agents, software that can carry out multi-step tasks on its own, perform cognitive work and robots increasingly perform physical work, we may have to rethink ownership, work, democracy, economic participation."

Big claim. But the smaller, practical version was already visible at TechBBQ. A software team losing access to an AI model overnight with no European alternative ready is not a hypothetical. It already happened.

For businesses building on tools they don't own, the lesson is to know your backup plan. Whittaker's point about convenience and privacy deserves equal attention from individuals: the two rarely travel together, and no one's asking you to sign a contract when you grant an AI access to your messages.

Common questions

What is AI sovereignty and why does it matter for Europe?

AI sovereignty means having local control over the AI models and computer infrastructure a country or region uses, rather than depending on foreign companies. Europe worries that if a US or Chinese AI provider restricts access or shuts down, European businesses have no fallback.

Why were Anthropic's Mythos and Fable models unavailable outside Europe?

The source material does not give a specific reason for the restriction. What it does confirm is that the cutoff was sudden, it disrupted at least some European software teams, and it focused minds on the risks of relying on foreign AI providers.

Should I worry about AI assistants built into my phone?

Meredith Whittaker's concern is that AI assistants integrated into messaging apps or operating systems can read large amounts of personal data. Her advice, in effect, is to think carefully before granting that access, especially when the trade-off is convenience rather than necessity.