What Did Stephen Hawking Think About Artificial Intelligence?

Stephen Hawking was one of the most prominent early voices urging caution about artificial intelligence. His position was not technophobia; it was a warning, from someone who relied on AI to speak, that we should think very hard about a technology that might one day surpass us.

The warning

In a widely reported 2014 interview, Hawking said that the development of full artificial intelligence could spell the end of the human race. His concern was specific: a sufficiently advanced AI might be able to redesign and improve itself, entering a cycle of self-improvement far faster than slow biological evolution can match. Humans, limited by the pace of natural change, could be left behind, or worse, find that a superintelligent system's goals did not align with our survival. He returned to the theme repeatedly and lent his name to open letters calling for research into keeping AI safe and beneficial.

The other side of the ledger

Hawking was careful never to be one-sided. He readily acknowledged that AI could be the most beneficial technology in human history, with the potential to cure disease, lift living standards and help solve problems beyond our current reach. He even depended on it personally: the system that gave him his voice used predictive software, a form of AI, to help him communicate. His message was not "stop," but "be careful, and start thinking about the risks now rather than later."

A prescient concern

There is a particular poignancy and authority in these warnings coming from Hawking. They were also, in hindsight, ahead of their time: the questions he raised about advanced AI, once treated as fringe speculation, have since moved to the centre of mainstream debate. How well his specific fears will age is assessed on the page about his predictions, but the basic instinct, to take a powerful emerging technology seriously before it arrives, is very much the Hawking approach, the same caution he urged about contacting aliens.