Anthropic is operating a lab that conducts biology experiments

Anthropic has confirmed that it has a bedewed biology laboratory wherever it tin usage its AI models to tally carnal experiments.

AI leaders person been promising that AI is the cardinal to curing quality disease. Dario Amodei opined conscionable past week: “I judge that AI could cure astir large diseases successful the adjacent 5–10 years.” To bash that, an LLM would person to person methods to trial its theories successful existent life.

“We judge that to bash biology, the last trial is still, and volition beryllium for a while, successful existent laboratory work,” Anthropic’s caput of beingness sciences, Eric Kauderer-Abrams told Reuters. “We perfectly are doing that today.” He added that the laboratory operates similar astir biotech labs: Anthropic conducts immoderate probe determination portion besides moving with externals partners.

This quality astir apt shouldn’t beryllium shocking. Anthropic bought Coefficient Bio, a stealth AI biotech, successful April. It declined to accidental what the bedewed laboratory is moving on, but that it’s not focused connected cause discovery. Anthropic doesn’t privation to springiness the quality of competing with the pharma industry, wherever it has galore large customers and partners. (For instance, it conscionable announced a woody to enactment with Novo Nordisk connected associated cause discovery.) Anthropic has already faced backlash for launching products perceived to vie with those of its customers.

To that end, Anthropic besides launched a Life Sciences Verification Program this week, to springiness vetted bio researchers access to its astir almighty models.

But the satellite is inactive reeling from the resignation of Anthropic researcher Jacob Coxon, who warned that “the radical gathering AI earnestly judge that it could termination america each by the extremity of the decade.” Anthropic’s ain alignment pb gave the likelihood astatine greater than 10% that AI could exterminate humanity wrong the adjacent decade.

The fervor has grown truthful aggravated that CEO Dario Amodei published a station past play calling connected the manufacture to dilatory down and institute self-regulation. He has repeatedly called bioterrorism 1 of AI’s biggest risks.

The juxtaposition of these dire warnings and the bedewed laboratory has not gone unnoticed successful the tech industry. As capitalist and AI coding startup founder Chamath Palihapitiya posted connected X, somewhat tongue-in-cheek: “The radical down specified hits as: “We’re All Going To Die” and “Regulate Me Now” are gathering a bedewed laboratory successful SF. I bash not urge this.”