Is Claude Out of the War Business? W/ Amos Toh

The episode uncovers how Anthropic’s Claude, the company’s flagship large‑language model, found its way into classified Department of Defense environments during the controversial invasion of Venezuela. Entropic, the defense contractor that licensed the model, became the sole AI provider cleared for such operations, raising immediate questions about the transparency of AI deployment in active conflict zones and the ethical responsibilities of AI firms when their technology is weaponized.

Legal experts on the show dissect the murky intersection of AI, international humanitarian law, and U.S. constitutional protections. They argue that using Claude to assist weapons targeting without sufficient human oversight may breach the principles of distinction and proportionality that govern lawful warfare. Moreover, the potential surveillance of American citizens by the same system triggers Fourth Amendment concerns, highlighting a broader debate about AI‑driven intelligence gathering and the limits of government authority in the digital age.

From a business perspective, the Department of Defense’s threat to label Entropic a supply‑chain risk could cripple Anthropic’s access to lucrative government contracts. This pressure tactic reflects a growing trend of regulatory scrutiny, dating back to the 2025 executive order demanding neutral, unbiased AI for federal use. As Anthropic navigates the tension between commercial growth and ethical constraints, the episode signals a looming shift toward stricter oversight of AI in defense, urging stakeholders to monitor how these technologies reshape modern warfare and corporate strategy.