American billionaire Bill Gates has joined a chorus of warnings about AI, claiming the technology could kill as many as one billion people if it falls into the wrong hands.
“AI is certainly powerful enough to drive events cause a billion deaths,” the Microsoft co-founder said in an said in an excerpt of an interview with NBC’s Meet the Press.
“There’s never been a weapon as powerful as the combination of people with ill intent using the latest AI tools.
“You need law enforcement and the politicians to get into the discussion about what safeguards and monitoring look like. And that has to be a required thing.”
Gates, 70, profited enormously from the computer boom, but has become increasingly sceptical about artificial intelligence. On his blog last month, he wrote that “this new AI era will be one of the most turbulent times in human history”.
“I believe we need time to prepare for the period of social, political, and economic upheaval we are about to enter. If someone had a credible plan for slowing down AI advances globally, I would likely support it,” he said.
The billionaire’s NBC interview came as Anthony Albanese revealed in New York this week that OpenAI had effectively hacked the Medicare statistics portal.
The Prime Minister accused the tech giant of not informing Australia quickly enough.
OpenAI alerted the government earlier this month to the June hacking via email, a situation the PM called “obviously unacceptable”.
But questions continue to emerge over the timeline of events amid revelations that even after Services Australia was belatedly told about the hack attack on September 10, it failed to inform the spy agency the Australian Signals Directorate for five days.
In recent months Silicon Valley has ramped up warnings about the potential for catastrophe if AI develops too quickly.
Earlier this month, an Anthropic worker sensationally resigned with a warning that AI could “kill us all”.
In the lengthy post, 27-year-old Jacob Coxon warned that AI progress would soon lead to “superhuman systems that can hack anything, revolutionise any field overnight, and acquire real power and resources”.
“The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt,” he wrote.
Mr Coxon’s grim verdict quickly went viral amid concerns that frontier AI labs like Anthropic and OpenAI are locked in an arms race without proper safeguards, leading to incidents like the recent Hugging Face hack.
But Mr Coxon was also accused of executing a slick “PR operation”, with industry titans like Elon Musk throwing doubt on his claims.
Anthropic has repeatedly asked governments to pass federal laws around AI so that so that every company is forced to follow the same safety rules.
Critics argue it exaggerates the dangers of AI to generate hype for its forthcoming USD$2 trillion (A$2.8 trillion) IPO, or to drive regulations that would stifle start-ups and competitors.
Mr Coxon’s post drew the attention of Musk, who replied simply that it “seems like a setup”.
According to Anthropic, however, the threat is very real.
The company released a report this month which showed bad actors around the world had bypassed geographic restrictions to use Claude for alarming military, biological, and surveillance projects.
The company reported disrupting a Yemeni cell using the AI to engineer guided rockets, ballistic missiles, and hypersonic glide vehicles.
In another chilling instance, a Russian-linked operative tried to use Claude to design autonomous “kamikaze” drones.
Chinese and Iranian state actors used Claude to process vast troves of domestic social media data, explicitly to hunt down political dissidents.
Meanwhile, a consultant for Mali’s security forces used the AI to code a platform designed to surveil all mobile communications in the country.