Although he has called DeepSeek China’s “best work”, Google’s DeepMind CEO Demis Hassabis has emphasised he sees “no actual scientific advance” with the AI model.
Google’s DeepMind CEO Demis Hassabis thinks DeepSeek is China’s “best work” – but has insisted there is “no actual new scientific advance” with the artificial intelligence (AI) model.
While the tech boss is impressed with the machine-learning model, Hassabis insisted “the hype” around DeekSeek has been overblown because the software is reliant on “known techniques”, many of which were developed by DeepMind.
According to Business Insider, Hassabis said yesterday (09.02.25) in Paris ahead of the AI Action Summit: “I mean, just briefly, on DeepSeek, it’s an impressive piece of work, and I think it’s probably the best work I’ve seen come out of China.
“But it’s important to understand that despite the hype, there’s no actual new scientific advance there. It’s using known techniques.
“Actually, many of the techniques we invented at Google and at DeepMind.”
Hassabis then pointed to Alpha Zero – DeepMind’s machine-learning software that managed to teach itself chess, go and shogi to the point where it could outclass world champions.
The DeepMind head said the Chinese AI model – which was built on a fraction of the cost of U.S. software like ChatGPT from OpenAI – could make a significant impact on the “geopolitical scale”.
When DeepSeek was introduced last month, it spooked investors in the U.S. AI industry – with chip juggernaut Nvidia seeing its market value come crashing down.
The software’s potential called into question the U.S.’s dominance in the AI field, with President Donald Trump describing DeepSeek as a “wake-up call” for American tech firms.
Google’s DeepMind CEO sees ‘no actual new scientific advance’ with China’s AI model DeepSeek
