Top AI agency Anthropic fears scientists used its chatbot to help bioweapon analysis
Leading AI firm Anthropic warns scientists used its Claude bot to assist sensitive virus research, raising fears the tech could be used to build bioweapons
Anthropic has warned that researchers may have used its chatbot to carry out bioweapon experiments (stock image)(Image: Getty Images/iStockphoto)
A leading artificial intelligence firm has issued a warning that researchers may have used its chatbot to carry out bioweapon experiments. Tech giant Anthropic said boffins used its smart bot, Claude, to help with sensitive biological studies that bosses fear could be hijacked for weapons development.
The work reportedly involved “gain-of-function” virus experiments and trials on pathogens potentially capable of unleashing a pandemic. Company chiefs admitted they couldn’t figure out the scientists’ true motives because the complex lab work can double as normal medical research or be weaponised to create super-bugs.
Anthropic slammed the brakes on the accounts and pulled the plug on their access after catching the policy breaches.
Anthropic said boffins used its smart bot, Claude, to help with sensitive biological studies(Image: NurPhoto via Getty Images)
The revelations were exposed in a report documenting how the artificial intelligence tech is being misused. It comes amid growing fears that rogue AI could wipe out humanity altogether.
Former Anthropic expert Jacob Coxon issued an apocalyptic warning earlier this week about bots, telling NBC News there was a “substantial probability that this technology could kill everyone.” He added: “This isn’t hyperbole. It’s not a marketing stunt.”
He announced his resignation from the company on X on Wednesday morning (September 9), alleging there is a lack of “responsibility” in the AI industry that is “gambling” with human life.
Former Anthropic expert Jacob Coxon issued an apocalyptic warning earlier this week about bots(Image: @hilbertspaess/X)
“I resigned from Anthropic today,” he said. “I spent the last three years doing pretraining research at both OpenAI and Anthropic.
“Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives.”
In one incident, a scientist reportedly asked Claude for help writing a grant application for research on the chikungunya virus, a painful mosquito-borne illness that causes severe fever.
Anthropic said the lab trial aimed to spot mutations that make the bug spread faster and dodge immune systems, before genetically engineering those changes straight into the virus.
While Anthropic acknowledged the research could technically help make life-saving vaccines, alarm was raised when they realised the project was targeted for a military research base.
The firm said it was unable to confirm if the military-backed bug work was being crafted as a bio-weapon, but took quick action to block the threat regardless.
