Washington, D.C., Jan 29: A major controversy has surfaced in the United States after reports claimed that sensitive government files were uploaded to the public version of ChatGPT. The allegations are directed at Madhu Gottumukkala, an Indian-origin official who served as the acting chief of the Cybersecurity and Infrastructure Security Agency (CISA) in May 2024.
According to Politico, several officials from the Department of Homeland Security (DHS) alleged that Gottumukkala pushed for access to AI chatbots soon after assuming charge, despite restrictions on their use within Homeland Security. Gottumukkala allegedly uploaded CISA-related contract files, some containing highly sensitive information, to ChatGPT’s public version. In August, CISA’s internal cybersecurity sensors flagged the uploads, warning of potential risks. DHS officials reportedly launched an inquiry into the matter, though it remains unclear whether the investigation has been completed.
CISA’s Public Affairs Director, Marcy McCarthy, stated that Gottumukkala had limited permission to use ChatGPT under DHS oversight. However, she acknowledged that any data uploaded to the public version would automatically be shared with OpenAI. This raises concerns that if users query ChatGPT, information linked to those files could potentially surface in responses.
The incident has sparked debate about the risks of using public AI platforms within sensitive government departments. Experts warn that even temporary permissions could expose classified or contract-related data to unintended audiences.
