AI will help make a Nobel prize-winning discovery within a year, says Anthropic co-founder | AI (artificial intelligence)

According to Anthropic’s co-founder, the AI ​​system will work with humans to make a Nobel Prize-winning discovery within 12 months, and bipedal robots will assist tradesmen within two years.

Jack Clark described a “dazzling sense of progress” in technology and made a number of predictions, including that companies run solely by AIs will generate millions of dollars in revenue within 18 months and that AI systems will be able to design their own successors by the end of 2028.

He also told a lecture at Oxford University on Wednesday that plausible scenarios remained in which the technology’s “chance of killing everyone on the planet is not zero” and that “it is important to make clear that this risk has not gone away.” Anthropic’s most popular model is Claude, but it recently released a version called Mythos that is proving alarmingly capable of exploiting cybersecurity weaknesses.

Clark told the students that it would be better if humans slowed down the development of technology, “giving ourselves more time as a species” to deal with the consequences of the forces of technology. But that won’t happen, he said, when there is a dizzying development “by a variety of actors and a variety of countries locked in competition with each other, where commercial and geopolitical rivalries often overwhelm the larger genre-specific existential aspects of the technology being built.” He said this was “not ideal.”

Clark is one of the most senior figures at Anthropic, which was founded by artificial intelligence researchers who left rival firm OpenAI over security disagreements. The $900bn (£670bn) company has been accused by Donald Trump’s White House and other AI accelerators of “fear mongering” to promote regulations that could strengthen its competitive position.

Anthropic disputes this, and Clark said many people are in denial about the advancement of artificial intelligence. He said he wanted to encourage humanity to prepare for a technology that “will soon be more collectively capable than all of us.” Comparing the failure to prepare for AI to the failure to prepare for pandemics such as Covid, he said: “If we wait and allow synthetic intelligence to proliferate, we will eventually be forced to be reactive.”

Critics of leading AI companies such as Anthropic, OpenAI and Google fear that over-reliance on a few AI models backed by large amounts of for-profit capital could create a “single point of failure” in global systems.

Professor Edward HarcourtThe director of the Institute for Artificial Intelligence Ethics, which hosted Clark’s conference, also warned that the rise of AIs that do more and more for humans risks creating “cognitive atrophy” that could weaken people’s decision-making and judgment. He has advocated alternative models of artificial intelligence, sometimes called “Socratic” artificial intelligence, that ask people to think more.

Clark said his most conservative prediction was that “a broad swath of the economy and society will undergo profound changes”; These may include the separation of the machine economy from the human economy, robots gaining brains, science advancing without humans, and scientific equipment that works but is not imagined by humans. He admitted that some of this sounded “crazy.”