US cyber defense chief accidentally uploaded secret government info to ChatGPT

Ars Technica
CISA's acting director, Madhu Gottumukkala, uploaded sensitive 'for official use only' documents to public ChatGPT, triggering internal security warnings.

Summary

Madhu Gottumukkala, the acting director of the Cybersecurity and Infrastructure Security Agency (CISA), accidentally uploaded sensitive, 'for official use only' contracting documents to the public version of ChatGPT last summer, according to reports from Department of Homeland Security (DHS) officials. This action triggered internal cybersecurity warnings designed to detect unauthorized disclosure of government material. Gottumukkala had sought and received special permission to use the chatbot, which most DHS staffers are blocked from accessing in favor of approved, secure AI tools. While the leaked information was not classified, it was sensitive enough that its exposure could adversely impact privacy or national interest programs. DHS investigated the incident, with potential consequences ranging from a warning to security clearance revocation, though CISA declined to confirm if the probe concluded. Gottumukkala's tenure has been marked by controversy, including mass layoffs that reduced CISA staff significantly, criticism from Congress regarding preparedness for election security threats, and a reported failure on a polygraph test, which he blamed on career staffers.

(Source:Ars Technica)