One of the founding figures in AI is warning of an artificial intelligence arms race following a recent panic over DeepSeek.
Yoshua Bengio, often called the 'godfather of AI,' has spent most of his career advancing artificial intelligence so he's probably one of the most trusted voices in the field.
Now, the Canadian machine learning pioneer is warning that the rush to build more powerful AI models could have serious consequences if it's not properly managed.
"The effort is going into who's going to win the race, rather than how do we make sure we are not going to build something that blows up in our face," Bengio said.
Advert
Bengio is also the author of the first International AI Safety Report - which is to be presented at an international AI summit in Paris next week.
Whether it be for economic or military purposes, the race for advanced AI technology can "result in cutting corners on ethics, cutting corners on responsibility and on safety," Bengio said, adding that it is "unavoidable."
He's in London currently with other AI pioneers to receive the Queen Elizabeth Prize for Engineering, a global award that recognises AI’s potential.
However, Bengio explained that his biggest worry is the uncertain consequences of the lack of AI regulation from President Trump and the race among tech giants to build ever more powerful AI models.
This arms race is causing systems to become 'more powerful' and almost 'superhuman in some dimensions,' he described.
Bengio stated: "As these systems become more powerful, they also become extraordinarily more valuable, economically speaking.
"So the magnitude of, 'wow, this is going to make me a lot of money' is motivating a lot of people. And of course, when you want to sell products, you don't want to talk about the risks."
However, not all the AI 'godfathers' agree.
Yann LeCun, Meta's chief AI scientist, is far less worried about AI spiralling out of control. In fact, he believes people are overestimating how smart AI really is.
"We have been deluded into thinking that large language models are intelligent, but really, they're not," he said. "We don't have machines that are nearly as smart as a house cat, in terms of understanding the physical world."
That said, LeCun does predict that AI will reach certain aspects of human-level intelligence within the next three to five years. By this, he means some robots could perform tasks without specific training or programming to do so.
Pointing the finger at Chinese chatbot company DeepSeek, LeCun said there's proof that no single country or company will dominate the AI race for long.
Releasing its first chatbot in January 2025, the company has already shown it can compete with the best of America's big tech using a fraction of the computing power.
"If the US decides to clam up when it comes to AI for geopolitical reasons, or, commercial reasons, then you'll have innovation someplace else in the world. DeepSeek showed that," he concluded.