'Godfather of AI' warns technology could wipe out humans within three years
Artificial intelligence could be capable of killing us within three years, according to one expert in the field.
This particularly bleak warning comes from none other than Professor Geoffrey Hinton, a Nobel Prize winner dubbed the 'Godfather of AI' due to his role in pioneering neural networks.
However, Hinton, like many other experts in the field of AI, has grown cautious about the technology's capabilities and is expected to outline the dangers humanity faces in an upcoming speech.
Delivering his remarks at the AI Safety Foundation's Hinton Lectures in Canada, Professor Hinton explained that he's revised his previous predictions about AI's capability after listening to lectures from Paul Christiano, founder of the Alignment Research Centre, who believed AI could kill us in as little as three years.
Advert
"I think for most people in this room, the most likely reason they will die is because of an AI accident," he said during Monday's opening lecture.
Professor Hinton, who is an unofficial AI advisor to the British Government, added in a second lecture on Wednesday: "I've been going around saying this stuff's going to get smarter than us and we really need to worry and it might happen within ten years. And so I'm the one who is scaring other people.
"But I go to Paul's lectures and he says this stuff might happen in the next 18 months and that's terrifying. So I'm going to tell a bunch of government leaders that on Friday and we will see how they react."
Explaining why his views on rapid AI development had changed, Professor Hinton added: "We’re now developing systems that will soon be much more intelligent than we are, and we have very little idea of how to ensure that they remain under our control.
"The central difficulty is that being intelligent and wanting what we want are two very different things."
This isn't the first time in which Professor Hinton has delivered sobering warnings on the development of AI, with the 78-year-old previously telling us to start being nice to intelligent software, ‘else we’re toast’.
Just in case you needed a reminder to say 'please' and 'thank you' to your chatbot next time you ask it a question.
Meanwhile, Christiano referenced the recent incident back in July where an OpenAI agent 'escaped' onto the internet and targeted Hugging Face software as an example of how quickly AI was developing.
"Last summer, you would have said that sounds kind of like crazy science fiction," he said, warning: "Future agents may end up treating humans as another obstacle."
Topics: Artificial Intelligence, AI, Technology
Brenna Cooper
