BY SRH
StevieRay Hansen Warns Death by Plague Is a Top AI Threat…
A group of AI CEOs demanded more regulations over the production of synthetic DNA earlier this summer. Stanford and Arc Institute researchers proved last month that AI can create new viral genomes. After a research by Anthropic revealed attempts to use Claude in ways that “could support biological weapons development,” the notion gained traction last week. “One of the most serious risks of frontier AI models” is biological misuse, according to the report’s writers. Not long after that, CEO Dario Amodei asked the government to assist AI research facilities in “pacing the frontier.”
On Tuesday, AI researcher Jacob Coxon announced his resignation from Anthropic and issued a dire warning that the AI arms race is endangering all of our lives. This news sent shockwaves through Silicon Valley and beyond. According to Coxon’s post on X, which has over 100 million views, many individuals working on AI agree with him that there is a pressing need to assure the safe development of AI systems before it’s too late.
According to Coxon, who was involved in the pretraining phase of AI development, “the consensus is that the next year or two is crunch time for humanity” (WIRED). Literal quotations from my coworkers at Anthropic are actually what I’m using here. They’ll use terms like ‘endgame’ and ‘crunch time,'” he explains. “When they look at it this way, Anthropic and its rivals get to choose humanity’s destiny.”
A group of AI CEOs demanded more regulations over the production of synthetic DNA earlier this summer. Stanford and Arc Institute researchers proved last month that AI can create new viral genomes. After a research by Anthropic revealed attempts to use Claude in ways that “could support biological weapons development,” the notion gained traction last week. “One of the most serious risks of frontier AI models” is biological misuse, according to the report’s writers. Not long after that, CEO Dario Amodei asked the government to assist AI research facilities in “pacing the frontier.”
An artificial intelligence researcher’s statement that there is a genuine possibility that AI might eradicate humanity within a decade caused his resignation from Anthropic to go viral on social media. Joined by Brian Barrett and Leah Feiger, WIRED’s Will Knight weighs in on whether the predictions of impending disaster are exaggerated or long overdue. We also talk about the new “always listening” Apple Watch capabilities that make Brian feel uneasy, as well as Apple’s $2,000 folding iPhone and the iPhone Duo. In addition, the Trump administration promoted a deceptive study that was based on inaccurate data, according to a WIRED investigation.
If you’re a Bluesky user, you can follow Brian Barrett (@brbarrett), Leah Feiger (@leahfeiger), and Will Knight (@willknight). Post your correspondence to uncannyvalley@wired.com.
