Trump's cyber chief accidentally uploads highly sensitive private documents to the public ChatGPT
President Trump's cyber chief has reportedly accidentally exposed private documents to ChatGPT.
While the use of artificial intelligence in our professional and personal lives is becoming less taboo, you've got to know where the line is drawn.
Some of us are guilty of sharing a bit too much with AI chatbots, whether it's financial information or personal data that probably shouldn't be fed into algorithms. Not to mention, tech leaders like Sam Altman are warning that messages to ChatGPT aren't all that 'private.'
But for one high-ranking US government official, things went sideways when 'sensitive' documents ended up somewhere they absolutely shouldn't have been.
Advert
According to a new report, the acting director of the Cybersecurity and Infrastructure Security Agency (CISA), Madhu Gottumukkala, uploaded contracting materials to the public version of ChatGPT.
The documents were marked 'for official use only,' which means they were sensitive and not for public release.
Serving under President Trump, Gottumukkala shared the files with the OpenAI platform last summer, unnamed Department of Homeland Security officials told Politico.
Their upload triggered automated security alerts designed to prevent the disclosure of this kind of leak, leading to an internal review by top DHS officials, the outlet said. Media reports say the findings of that review haven't yet been made public.
According to three officials (via Politico), Gottumukkala had pushed for special access to ChatGPT shortly after joining CISA in May.
At the time, the AI tool was off-limits to nearly all DHS staff due to worries that sensitive information could be stored or exploited outside federal systems.
Yet the former chief information officer for South Dakota's Bureau of Information and Technology managed to secure special access by essentially 'forcing CISA’s hand into making them give him ChatGPT,' one official told Politico. “And then he abused it.”
The problem is that anything entered into ChatGPT's public version can be incorporated into future prompts, meaning the files could be exposed to as many as one billion users worldwide.
In an emailed statement to Politico, Marci McCarthy, CISA’s director of public affairs, seemed to downplay the incident. She confirmed that Gottumukkala received 'permission to use ChatGPT with DHS controls in place,' and described his use as 'short-term and limited.'
She wrote: “Acting Director Dr. Madhu Gottumukkala last used ChatGPT in mid-July 2025 under an authorized temporary exception granted to some employees.
“CISA’s security posture remains to block access to ChatGPT by default unless granted an exception.”
McCarthy also emphasised the agency's priorities on advancing America's leadership in AI, following Trump's January 2025 executive order on the subject.
Gottumukkala has been serving as CISA's director since May, after DHS Secretary Kristi Noem appointed him as deputy director. His time in the position has already been marked by multiple scandals, including reportedly failing a polygraph examination that he himself had initiated, Politico reported.