Skip to content

AI Pioneer Aidan Gomez Warns Democracies Are Losing the AI Race to China—And Europe Must Act Now

At a G7 lunch, on a VivaTech stage, and over dinner with reporters, the Cohere CEO and Transformer Co-Creator made his pitch for greater urgency and resources to build sovereign AI. Before it's too late.

Aidan Gomez had just come from a room that, by his own account, almost never exists. Leaders of the world's biggest democracies were at the table for a G7 lunch, and so were the people who run the AI labs.

"You don't get them with the AI CEOs pretty much ever," the Cohere co-founder and chief executive said on stage at VivaTech. What surprised him was a level of agreement he hadn't anticipated: a shared sense that democracies need to deploy this technology faster and coordinate on the downsides, naming child safety and the cyber threat as the obvious candidates. Yes, even the U.S., he insisted the next day to a roomful of skeptical journalists.

His contribution to the G7lunch, he later recounted, focused on AI resilience and democracy. While America has led the spending and built some of the most capable AI systems, he said that the country sitting in second place in terms of the most powerful LLMs should be a democracy. And right now it isn't.

"If we want democracy to continue to be the dominant system of governance," he said, multiple democracies have to hold the top positions, not just the United States holding the first position.

That is central to the mission that drives his own company, Toronto-based Cohere, which focuses on building models for enterprises. And even as Gomez pushes that business ahead, his appearance in Europe last week had the feel of a man leading a campaign and looking for new recruits to his cause.

In the beginning

Cohere was founded in 2019 by Gomez, Ivan Zhang, and Nick Frosst. Gomez and Frosst were part of the research team at Google Brain that developed the transformer machine learning architecture described in the landmark 2017 paper "Attention Is All You Need." This architecture subsequently became the foundation for creating large language models such as ChatGPT.

Become a free member to read this post

Subscribe

Already have an account? Sign In

Latest