Elon Musk has been sounding the alarm over the years over the potentially dangerous, artificial future of artificial intelligence.
In 2016, the billionaire stated that humans could be the equivalent of "house cats" for the new AI overlords. Since then, he has repeatedly called for regulation and caution when it comes to new artificial intelligence technology.
But of all the various AI projects in development, none of Musk cared more than Google's DeepMind.
“It is the nature of the AI that they create that overwhelms all people in all games,” Musk told The New York Times. "I mean, it's basically a storyline in WarGames."
In WarGames, a teenage hacker, played by Matthew Broderick, connects to an AI-controlled government supercomputer trained to run military simulators. In an attempt to play a game called Global Thermonuclear War, the AI convinces government officials that a nuclear attack from the Soviet Union is imminent.
At the end (spoiler for those who haven't seen the 37-year-old movie), the computer runs enough simulations of the end results of a global thermonuclear war that it announces that there is no possible winner, and the only way to win is not to play. The 1983 film is a direct reflection of its time and place, when US fears of a nuclear war with the Soviet Union continued to loom alongside fears of increasingly advanced technology.
But Musk wasn't just talking about old movies when he compared DeepMind to WarGames - he also said that AI could surpass human intelligence in the next five years, even if we don't see its impact immediately.
“This does not mean that in five years everything will go to hell,” he said. "It just means things are getting unstable or weird."
Musk was reportedly one of the earliest investors in DeepMind, which was sold to Google in 2014 for $ 500 million. In a 2017 interview, he said he took the step to follow the explosive development of AI, and not to return on investment.
“It gave me a better understanding of the rate at which things were improving, and I think they are indeed improving at an accelerating rate, much faster than people think,” he said in a 2017 interview. “Mainly because in everyday life you don't see robots walking. Maybe your Roomba or something. But Roombas is not going to take over the world."
But Musk thinks artificial intelligence should have a different meaning.
“I think, in general, people underestimate the power of AI - they think it's a smart person,” Musk said in August speaking with Alibaba CEO Jack Ma at the World Conference on Artificial Intelligence in Shanghai. “But it will be much more than that. It will be much smarter than the smartest person."
It is “pride,” he told The Times this week, that prevents “very smart people” from realizing the dangers of AI.
“My assessment of why AI is overlooked by very smart people is that very smart people don't think a computer can be as smart as they are,” he said. "And this is pride and obviously a lie."