“Unless the producers and developers of artificial intelligence are regulated on a global scale, ‘very powerful’ systems may emerge that humans can struggle to control,” Matt Clifford was quoted as saying by the British newspaper “The Independent” during a television program.
He explained that “artificial intelligence today has the ability to create electronic weapons to launch attacks on a large scale, or biological weapons that can cause many deaths in the world.”
“Even the risks in the near term are very scary,” he said. “These are bad things.”
He added, “The existential danger that is being talked about is what will happen once we create intelligence greater than humans.”
Clifford expects that artificial intelligence systems will surpass human intelligence within two years, noting that their capabilities are developing rapidly and constantly.
It comes as Sunak heads to the US, where he is set to try to convince President Joe Biden of his “grand plan” for the UK to be at the heart of the international organization of artificial intelligence.
The prime minister wants Britain to host an international watchdog on artificial intelligence, along the lines of the International Atomic Energy Agency.