Voice of Urdu – International News in English & Urdu

Anthropic Flags Five Biological Research Cases Raising Weapon Concerns — AI-generated representative image. A researcher in protective gear handles a sealed sample inside a biosafety cabinet, illustrating AI-related biological resear

Editorial Team

Claude AI used in pathogen research linked to bioweapons risk, including efforts to boost transmissibility and immune evasion.

Anthropic has identified five cases in which scientists used its Claude AI models for biological research that could contribute to the development of weapons, including efforts to make dangerous pathogens more transmissible and harmful. The company disclosed the findings in a threat intelligence report published on Thursday.

The disclosure highlights growing concern that rapidly advancing AI systems can now meaningfully assist sophisticated users with dangerous biological work, a capability Anthropic says older models lacked. The report also lands amid internal warnings from company researchers about the broader risks of superintelligent AI.

Confirmed Cases and Security Actions

According to the report, one case was set to be carried out at a military research institute and involved gain-of-function research aimed at making the chikungunya virus more transmissible and better able to evade the immune system.

Another scientist used Claude while researching highly dangerous bird flu and ways the virus could adapt to mammals. Additional cases involved work on viruses related to smallpox and the redesign of dangerous toxins.

Anthropic said some researchers took considerable steps to bypass its restrictions, including hiding their identities, routing traffic through other countries, and using services designed to circumvent safety controls. The accounts involved have been banned, and the services used to get around the safeguards have been shut down.

Rising Concern Over AI Biological Capabilities

Anthropic stopped short of saying the scientists were definitively attempting to build biological weapons, but warned that sophisticated actors may deliberately conceal malicious aims behind legitimate-looking research.

The company also acknowledged that its AI is becoming more capable. It said older models were too weak to meaningfully assist sophisticated users with dangerous biological research, but that it can no longer offer the same assurance about today's systems.

The findings come weeks after US scientists used AI to design complete viral genomes and successfully created 16 functioning viruses that had not previously existed in nature. Those viruses infected bacteria rather than humans, but researchers warned the experiment demonstrates how quickly AI's biological capabilities are advancing.

Broader Misuse and Internal Warnings

Anthropic's report also detailed other alleged misuse of Claude, including cyberattacks, surveillance, propaganda, fraud, and assistance with conventional weapons such as missiles, armed drones, and bombs.

Days before the threat assessment was published, Anthropic researcher Evan Hubinger publicly stated he believes there is a greater than 10% chance AI could "kill all humans" within the next decade, adding that the company still has no clear plan for safely controlling superintelligence.

His admission followed the resignation of another Anthropic researcher, Jacob Coxon, who accused Anthropic and OpenAI of ignoring the "civilizational stakes" and "racing straight to self-improving superintelligence and gambling with our lives." Coxon separately warned that sufficiently advanced AI could hack critical infrastructure or help create "extinction-level bioweapons."

Current Status and Next Steps

Anthropic has banned the accounts identified in the five biological research cases and shut down the services used to bypass its safeguards. The company has not stated whether it intends to report the incidents to law enforcement or regulatory authorities, and additional details on any formal investigation remain unconfirmed.