DeepSeek Is Great But Hype Is ‘exaggerated,’ DeepMind CEO Thinks

Chinese AI company DeepSeek continues to generate a lot of opinions among experts of all stripes. Some have focused on the potential national security and privacy risks of its AI models. However, others have praised the firm’s engineering prowess. Recently, a top Google DeepMind executive praised DeepSeek AI, but he also believes the hype is overblown.

Today, the Paris AI Action Summit, a two-day conference dedicated to advances in the segment, will begin. In the days leading up to it, Google reportedly held an event where Demis Hassabis, Nobel laureate and head of its AI division DeepMind, offered a few words about DeepSeek.

Google DeepMind CEO praises DeepSeek AI, but thinks the hype around it has been “exaggerated”

Hassabis reportedly said that DeepSeek AI models are “an impressive piece of work.” “I think it’s probably the best work I’ve seen come out of China,” he added. Hassabis believes that much of the company’s success is due to “extremely good engineering” and that it changed things “on a geopolitical scale.” Recall that the availability of DeekSeek’s R1 and V3 models caused crashes on Wall Street for companies as large as NVIDIA. However, lately some experts have been questioning the firm’s alleged low investment.

That said, Hassabis also believes that the hype around the AI ​​platform has been “exaggerated a little bit.” This is because the company’s models do not actually bring anything new in terms of functionality to the industry. “Despite the hype, there’s no actual new scientific advance…it’s using known techniques [in AI],” he stated.

The executive says that the future of the AI ​​industry is artificial general intelligence (AGI). That is, the “AI agents” that can perform various tasks as well as a human would. “I think we’re close now, you know, maybe we are only, you know, perhaps five years or something away from a system like that which would be extraordinarily pretty,” Hassabis claimed.

Lastly, Hassabis also gave his opinion on regulation in the artificial intelligence sector. “It’s important to regulate AI but it’s important to get the regulations right. That’s hard when the technology itself is not fully understood and so fast-moving… There needs to be international cooperation around that, which is also complicated in the current environment,” he said.