Amid stark warnings that artificial intelligence could wipe out humanity by the end of the decade, executives from the largest artificial intelligence companies in the United States have backed Anthropic CEO Dario Amodei’s bid to slow development to ensure public safety. Anthropic co-founder Jack Clark went further, saying that all artificial intelligence software should have a mandatory “kill switch” so that society can “turn off” anything that becomes too dangerous.
The warnings sent both tech stocks and Donald Trump into a tailspin. The US President has called calls to “limit” the development of artificial intelligence a “SICK conspiracy” that will only benefit China.
We’ve long since moved past the “more mundane” warnings that AI is leading to job losses or cheating in exams, said BBC Tech and AI editor Zoe Kleinman. Talk of “existential threats” to humanity has led to “a chorus of serious voices calling for a slowdown.” Yet it’s “far from a simple solution” because it relies on artificial intelligence companies that “have not behaved responsibly enough to date.” It’s like a bull running out of a farmer’s field and wreaking havoc, leading computer scientist Wendy Hall told Kleinman. “Of course it’s not the bull’s fault, it’s the farmer’s.”
Week
Escape from your echo chamber. Get the facts behind the news, plus analysis from multiple perspectives.
SUBSCRIBE AND SAVE
Sign up for the free newsletter of the week
Join over 350,000 subscribers and stay up to date with the week’s most interesting, educational and entertaining stories, as well as daily puzzles.
Join over 350,000 subscribers and stay up to date with the week’s most interesting, educational and entertaining stories, as well as daily puzzles.
Latest videos fromWeek
The problem, according to AI skeptic Ed Zitron, is that “no one has given a substantive explanation of what ‘slowdown’ means.”
Amodei’s broad offerings include independent monitoring of AI models as they are developed, as well as global agreements on safety standards and security issues. While this is commendable, many believe it is unrealistic given the current geopolitical landscape. “The US federal government is committed to a global artificial intelligence arms race,” and “previous proposed AI legislation has weakened under intense and well-funded corporate lobbying,” The Guardian reports.
And then there is China. The state-run Global Times described Amodei’s essay as a thinly disguised attempt to curb Beijing’s technological ambitions, saying it was “filled with containment provisions aimed at China and essentially presents a ‘Cold War scenario’ for the artificial intelligence sector.”
America “appears to be increasingly weaponizing its dominance in the sector, viewing it as a strategic asset that must be protected,” said CNN’s John Liu. China, which is “behind the race,” favors a “more open model” but has given no indication that it will reflect any slowdown in artificial intelligence development in the US. With so much at stake, it seems neither side is willing to back down.
A free daily email newsletter with the day’s biggest news and analysis, plus TheWeek.com’s best features.
What’s next?
Anthropic found itself in a “classic prisoner’s dilemma,” said Tim Higgins in The Wall Street Journal. If he is “the only one to heed the call, he will fall behind OpenAI”; if all American artificial intelligence labs achieve a slowdown, “it may well benefit all of humanity,” but it would likely cause America to “fall behind China.”
With both Anthropic and OpenAI planning to list shares on the US stock market, some are also wondering whether their newly discovered security issues are what they seem. “This is the story of the IPO in essay form: We are a responsible firm; we are the medicine; buy the security premium,” said AI entrepreneur Brian Roemmele on X. For him, these dire warnings are nothing more than a sales pitch.
