After US AI Ban, Swedish Researcher Urges Europe to Invest in Own Models

The U.S. government's decision to block non-Americans from accessing Anthropic's advanced AI tool, Mythos, has sparked concern among European researchers. Pontus Johnson, a professor at KTH Royal Institute of Technology, argues that Europe must significantly increase investment in its own AI language models to avoid falling behind in global competitiveness.

Mythos, developed by the U.S. AI company Anthropic, is reportedly 10 to 100 times more powerful than previous tools for identifying and exploiting software vulnerabilities. While Anthropic acknowledges the potential risks of the tool falling into the wrong hands, Johnson views the U.S. government's restriction as a potential precursor to broader limitations on Europe's access to cutting-edge AI technology.

“

It may be that the U.S. state, which has been very permissive so far, begins to tighten the screws. It may become more difficult for Europe to gain access to this very important technology.

Johnson likens the situation to the control surrounding nuclear technology, suggesting that as AI becomes more powerful, governments will increasingly treat it as a national security issue. He advocates for substantial investment in European AI models, pointing to the French company Mistral as a potential foundation for this effort. Failure to do so, he warns, could severely impact Europe's technological standing.

“

It is completely obvious that Europe must have its own large language models. We have no one that measures up to the American models today. The best we have is Mistral. I think the best way for us is to invest heavily in it.