Many of them didn’t hack in. They signed up.

By Jeff Newman Esq. MBA

On September 10, Anthropic published a threat report on misuse of its Claude AI model that it detected and blocked. Among the cases:

→ An actor linked by account information to the PLA’s Academy of Military Sciences modeled radar jamming against 12 targets in Taiwan, including Patriot and Tien Kung batteries.

→ A Yemen-based cell used Claude on guidance software for a rocket that was later test-fired.

→ A Russia-based procurement manager used it to find Chinese and Hong Kong intermediaries for export-controlled European goods.

There is one word in the report that deserves more attention. Anthropic says operators “laundered” their access through VPNs, foreign phone numbers, fraudulent accounts, stolen credit cards and stolen API keys.

That word matters legally. Under Van Buren v. United States, breaching terms of service is probably not a federal crime. Funding accounts with stolen cards is. Running workloads on keys taken from an American company is.

The state actors in the headlines are beyond reach. The proxy operators, card processors, key brokers and resellers in the middle may not be.

Full analysis, including the OFAC reporting question and the FinCEN and DOJ whistleblower programs, in the first comment. #Whistleblower #Sanctions #ExportControls #AI #Compliance

Jeff Newman Law is a national whistleblower law firm that handles False Claims Act cases, SEC whistleblower cases, CFTC whistleblower cases, IRS and more. The firm website is www.JeffNewmanLaw.com

#Whistleblower #Sanctions #Exportcontrols