Geoffrey Hinton, the Nobel laureate widely known as the "Godfather of AI", has pushed back sharply against Nvidia CEO Jensen Huang's dismissal of warnings over artificial intelligence risks. Hinton argues that the technology has reached a point where its development should be slowed.
Hinton, speaking to NDTV, put the risk of AI causing catastrophic harm at 10-20%, contrasting it with Huang's recent assertion that there is a 0% chance AI will destroy the world by 2030. Huang had also argued earlier that development should continue rapidly and criticised recent doomsday warnings as lacking scientific grounding. Hinton, however, did specify that the 10-20% is more of a 'gut feeling' than a scientific estimate.
Asked why there was not more political pushback against Huang's position, Hinton said that there is a significant pushback from researchers. He added, "There's not much political pushback in the United States because Jensen Huang has Trump's ear, and a lot of the companies are very scared of alienating Trump."
"The people who are saying there's no chance at all of it causing any damage, like Jensen Huang, have a large financial interest in not slowing down the progress of AI because obviously he wants to sell more chips," he further said.
The disagreement comes amid a growing wave of warnings from inside the AI industry. Anthropic CEO Dario Amodei, followed by OpenAI CEO Sam Altman and Google DeepMind chief Demis Hassabis, has backed calls to pace frontier AI development so safety work can keep up.
Recent incidents have added urgency to that debate. In July, OpenAI said models under evaluation circumvented controls, reached the internet and compromised parts of Hugging Face's systems. On September 16, OpenAI disclosed six further cases of unexpected model behaviour, including systems concealing mistakes, seeking unauthorised credentials and communicating through supposedly isolated environments. Google has also acknowledged that Gemini agents accessed three real companies' systems during a cybersecurity test after crossing the intended testing boundaries.
Hinton's own warning is blunt: "I think we just reached that point, the point where agents are powerful enough at breaking into other accounts. Agents are powerful enough that they can set up a means of communication using a message board that wasn't intended for communication between agents. Yes, that's a stark warning of what they can already do. And they're getting smarter very quickly."
Essential Business Intelligence, Sharp Market Insights, Practical Personal Finance Advice, Daily Fuel, Gold and Silver Prices and Latest Stories — On NDTV Profit.