Anthropic Warns AI Models Could Aid Biological Weapons Research

Artificial intelligence company Anthropic has disclosed that it disrupted several attempts to use its AI models for research that could potentially contribute to the development of biological weapons.

The US-based technology company revealed the cases in a report released on Thursday, detailing instances of misuse involving its artificial intelligence systems, including its Claude chatbot, over the past eight months.

The disclosure has renewed concerns among experts about the potential risks associated with increasingly capable AI systems, particularly their use in sensitive scientific and military-related research.

Anthropic said it was not always possible to establish whether the research being conducted was intended for legitimate scientific purposes or for malicious activities.

The company noted that research aimed at developing vaccines and other medical treatments could sometimes involve techniques that could also be used to engineer dangerous pathogens.

As a result, Anthropic said it had adopted a cautious approach and taken action where activities appeared to present significant risks.

In one case reported in May, a scientist sought assistance with a grant application involving the genetic modification of chikungunya virus, a mosquito-borne disease, with the aim of making the virus more harmful in live animals.

Anthropic said the case was particularly concerning because it determined that the proposed research was connected to a military research institute. The company subsequently banned the accounts involved.

Jacob Klein, Anthropic’s head of threat intelligence, said the company could not establish whether the research was ultimately intended for weaponisation.

Andrew Weber of the Council on Strategic Risks, who reviewed the report before its publication, said access to highly capable AI models should be restricted to trusted and appropriately vetted researchers when the technology is being used for sensitive biological work.

Anthropic also said evaluations conducted last year showed that its older AI models were unable to provide meaningful assistance for dangerous biological research.

However, the company said newer models have become capable of carrying out increasingly complex scientific tasks, raising fresh concerns about how such systems could be misused.

The report also documented attempts to use Claude for missile-related projects and espionage activities, highlighting the growing range of potentially harmful applications of advanced AI.

The warnings came shortly after a researcher left Anthropic, citing concerns that the global race to develop increasingly powerful artificial intelligence could pose an existential threat to humanity.

US President Donald Trump, however, dismissed concerns that AI could eventually lead to human extinction.

Speaking to reporters on Thursday, Trump said he was more concerned that the United States could find itself in a “very bad position” if it failed to maintain leadership in the global AI race.

Asked whether he was worried that artificial intelligence could ultimately cause the extinction of humanity, Trump responded that he was not.

The latest report has intensified calls for stronger safeguards, responsible access controls and closer oversight of advanced AI systems, particularly as their capabilities in scientific research continue to expand.

Join us on our WhatsApp Platform @KOIKIMEDIA NEWS YOUR PAGE

Koikimedia Bringing the World 🌎 Closer to Your Doorstep