Anthropic, the AI tech company behind Claude, released a report Thursday detailing how bad actors have attempted to use its products for nefarious ends, from scams and mass surveillance to biological weapons development.
“We’re publishing this work because we believe we have a responsibility to disclose malicious misuse of our services,” the San Francisco company said in its 154-page report. “As models become increasingly capable, their risks will increase, unless AI developers and society’s defenders act to make them safer.”
For example, last May, Anthropic’s biological safety classifier blocked a request for Claude’s assistance in authoring a grant application for scientific funding. The reason was that the project was for research into the highly contagious virus known as chikungunya. So-called gain-of-function research can theoretically make viruses more transmissible or more deadly. “Similar research could certainly be used in the development of better vaccines and therapeutics for the virus — but it could also be used to make the pathogen more dangerous,” Anthropic wrote. To make matters more suspicious, the proposed research facility was a military research institute.
*Start your day with essential news from Salon.
Sign up for our free morning newsletter, Crash Course.* Even though Anthropic has released similar threat assessment reports in the past, the timing for this one comes just days after Jacob Coxon, an Anthropic researcher, resigned from the company, claiming they were not acting responsibly with AI development.
“Do not underestimate the power of this technology,” Coxon said on X. “These will soon be superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources. We have all witnessed the progress in each of these domains, and progress is not slowing.”