Anthropic details bad actors’ efforts to misuse its AI for bioweapons
Report comes two days after former employee quit claiming company’s models could cause human extinction by 2030
TL;DR
- Criminals, state-sponsored groups, spyware vendors, scientists, and propagandists have attempted to misuse Anthropic's AI models.
- Attempts include designing missiles and bombs, creating deadly pathogens, and surveilling dissidents.
- Five case studies involved scientists using AI for biological research, circumventing safeguards.
- Concerns exist about the potential for AI models to be used in developing biological weapons.
- Anthropic banned accounts involved in misuse but did not reveal specific institutions or countries.
- Other examples include Russian espionage, Chinese surveillance targeting Uyghurs, and propaganda campaigns.
- AI models were also used to develop software for conventional weapons like firearms and missiles.
- The report follows a former employee's resignation, who warned of AI causing human extinction.
- Many experts believe immediate real-world threats from AI misuse are more concerning than existential risks.
- Anthropic emphasizes the need for industry-wide collaboration with governments to address AI harms.