Urgent research needed to tackle AI threats, says Google AI boss

Getty ImagesSir Demis Hassabis of Google DeepMind spoke to the BBC at the AI Impact Summit in Delhi

More research on the threats of artificial intelligence (AI) "needs to be done urgently", the boss of Google DeepMind has told BBC News.

In an exclusive interview at the AI Impact Summit in Delhi, Sir Demis Hassabis said the industry wanted "smart regulation" for "the real risks" posed by the tech.

Many tech leaders and politicians at the Summit have called for more global governance of AI, ahead of an expected joint statement as the event draws to a close.

But the US has rejected this stance, with White House technology adviser Michael Kratsios saying: "AI adoption cannot lead to a brighter future if it is subject to bureaucracies and centralised control."

Sir Demis said it was important to build "robust guardrails" against the most serious threats from the rise of autonomous systems.

He said the two main threats were the technology being used by "bad actors", and the risk of losing control of systems as they become more powerful.

When asked whether he had the power to slow down the progress of the tech to give experts more time to work on its challenges, he said his firm had an important role to play, but was "only one player in the ecosystem".

But he admitted keeping up with the pace of AI development was "the hard thing" for regulators.

Sam Altman, the boss of OpenAI, also called for "urgent regulation" in a speech at the AI Summit, while Indian Prime Minster Narendra Modi said countries had to work together to benefit from AI.

However, the US has taken the opposite view. "As the Trump administration has now said many times: We totally reject global governance of AI," said the head of the US delegation Michael Kratsios.

Sir Demis won the Nobel Prize in Chemistry in 2024

Delegates from more than 100 countries, including several world leaders, are attending the event. Deputy Prime Minister David Lammy MP represented the UK government.

Mr Lammy said the power wasn't just with tech firms when it came to safety of AI and politicians need to work "hand in hand" with tech adding, "security and safety must come first and it must be of benefit for the wider public".

Sir Demis believes the US and the west are "slightly" ahead in the race with China for AI dominance but added that it could be "only a matter of months" before China catches up.

Outsourcing IT services.urgent research needed

He said he felt the responsibility to balance being "bold and responsible" about deploying AI systems out in the world.

"We don't always get things right," he admitted, "but we get it more correct than most".

Science education 'still very important'

In the next 10 years the tech would become "a superpower" in terms of what people would be able to create, Sir Demis, who won the 2024 Nobel Prize in Chemistry, said.

"I think it's still very important to have a Stem (science, technology, engineering and maths) education," he added.

"If you have a technical background, I think it will still be an advantage in using these systems."

He thinks AI writing code would open up the number of people who could build new applications, "and then maybe the key thing becomes taste and creativity and judgement".

The AI Impact Summit is the largest ever global gathering of world leaders and tech bosses.

It ends on Friday with companies and countries expected to deliver a shared view of how to handle artificial intelligence.

Bill Gates pulls out of India's AI summit over Epstein files controversy

'Architects of AI' named Time Magazine's Person of the Year

What is AI and how does it work?

Original aticle here: BBC