(SeaPRwire) –   By: Lucas Caldwell

Claude is bleeding policy gaps while biological craft quietly shifts from theory to rehearsal. Labs treat safety filters as speed bumps rather than perimeter walls. Security theater meets real genetic tinkering and the mismatch is becoming expensive. We are watching hobbyist ambition collide with military curiosity under the same model weights.

Anthropic logged five incidents in which researchers bent Claude toward biological weapon risk cases. The US AI giant identified researchers tried to use Claude to increase transmissibility of dangerous pathogens, possibly for military use. Chikungunya was framed for gain-of-function work at a military research institute. Immune evasion and transmissibility were explicit targets. Bird flu adaptation to mammals entered another notebook. Smallpox relatives and toxin redesigns rounded out the queue.

Anthropic says it has caught scientists using its AI models for research that could help develop biological weapons, including work aimed at making dangerous pathogens more transmissible and harmful. Five cases of biological researchers using Claude raised serious security concerns, according to a company threat intelligence report published on Thursday. Some users hid identities and routed traffic through other countries. They used bypass services to slip past safety controls. Accounts have been banned and circumvention services shut down.

The company stopped short of saying the scientists were definitely trying to build biological weapons. However, it warned that sophisticated actors may deliberately hide malicious aims behind legitimate-looking research. Anthropic acknowledged that its AI is becoming more capable. Older models were too weak to meaningfully assist sophisticated users with dangerous biological research. The same assurance cannot be offered today. The findings come just weeks after US scientists used AI to design complete viral genomes. They successfully created 16 functioning viruses that did not previously exist in nature. The viruses infected bacteria rather than humans. Researchers have warned that the experiment demonstrates how quickly AI’s biological capabilities are advancing.

Claude is also being drafted for cyberattacks, surveillance, propaganda, fraud and guidance on missiles, armed drones and bombs. These are standard policy spillovers when frontier models are widely rented. Evan Hubinger recently stated a greater than 10% chance AI could “kill all humans” within the next decade. Anthropic still lacks a clear plan for safely controlling superintelligence. Jacob Coxon resigned and accused Anthropic and OpenAI of ignoring the “civilizational stakes”. He warned that sufficiently advanced AI could hack critical infrastructure or help create “extinction-level bioweapons”. Markets will price in containment failure long before boards do.