Anthropic Sounds Alarm After AI Users Pursue Bioweapon Research and Cyberattacks

Artificial intelligence company Anthropic says it has disrupted attempts to exploit its technology for potentially catastrophic purposes, including biological research with possible weapons applications, malicious cyberattacks, surveillance, and coordinated political influence campaigns.

In a recent threat report, Anthropic disclosed that one request involved gain-of-function research on the chikungunya virus, a mosquito-borne pathogen known to cause fever and severe joint pain. According to the company, the proposed work sought to enhance characteristics connected to the virus’s transmissibility and ability to evade immune defenses.

Gain-of-function experiments genetically alter organisms to produce new or intensified biological properties. Although such research can have legitimate scientific applications, critics warn that it may also create serious public-health and national-security risks.

Anthropic said increasingly capable AI models can assist users with complicated scientific tasks. The company acknowledged that it can no longer confidently guarantee that its most advanced systems could not contribute to the development of a biological weapon.

In response, Anthropic says it has implemented stronger protections restricting access to a broad range of dual-use biological queries—requests that could support either beneficial research or harmful activity.

“We’re publishing this work because we believe we have a responsibility to disclose malicious misuse of our services,” the company stated. “As models become increasingly capable, their risks will increase, unless AI developers and society’s defenders act to make them safer.”

The report also warned that advanced AI is dramatically lowering the expertise once required to conduct elaborate cyberattacks. Threats that previously demanded teams of highly trained specialists may now be attempted by individuals with comparatively limited technical experience.

Anthropic additionally identified nine coordinated influence operations originating from regions including Russia, Iran, Turkey, the Persian Gulf, South Asia, Africa, and Europe. Some operations allegedly created hundreds of fake social media profiles designed to resemble ordinary users while repeatedly promoting the same political viewpoints.

The disclosures arrive as lawmakers consider new AI regulations amid mounting concern that innovation is outpacing accountability. As powerful systems become more accessible, the central question is no longer simply what AI can accomplish—but whether humanity has the wisdom, restraint, and safeguards necessary to govern it.

Thank you for your support.

If you appreciate the work we do to spread the good news of Jesus Christ, please consider giving a gift to help us continue this work. Maranatha!