The US government website used a Chinese AI model

The official portal of Federal Register temporarily integrated a Chinese-origin AI search tool, which drew criticism from law enforcement as the FBI had previously classified the company's developments as malicious.

According to Reuters, the incident is related to accusations against several Chinese tech firms, with regulators claiming they aggressively copied the architectures of leading American models, including those developed by Anthropic. The use of third-party code on the government resource occurred without proper security checks.

During the investigation, it was found that the search engine was embedded in the site's infrastructure to process user requests. The FBI had repeatedly warned about the risks of data leaks when working with models that do not meet US cybersecurity standards. The integration has currently been disabled, and the agency is conducting an audit of all external APIs used.

This case highlights the vulnerability of government systems when implementing third-party AI solutions. Experts note that using models with opaque origins poses a threat to intellectual property and data security. The US government plans to tighten verification protocols for all tools integrated into federal web services.