US cyber defense chief accidentally uploaded secret government info to ChatGPT

US cyber defense chief accidentally uploaded secret government info to ChatGPT

In a troubling incident, the acting head of the Cybersecurity and Infrastructure Security Agency (CISA), Madhu Gottumukkala, mistakenly uploaded sensitive government documents to a public version of ChatGPT last summer, according to reports from Politico. Four officials from the Department of Homeland Security (DHS) familiar with the situation revealed that this incident triggered several internal cybersecurity alerts aimed at preventing unauthorized disclosure of government materials. The incident occurred shortly after Gottumukkala's appointment, during which he sought special permission to access OpenAI's widely used chatbot—access that is typically restricted for most DHS personnel. Instead, DHS employees are expected to utilize approved AI tools, such as DHSChat, which are specifically designed to keep sensitive queries and documents within federal networks. The reasons behind Gottumukkala's need to use ChatGPT remain unclear. One insider suggested that it appeared he pressured CISA into granting him access, which he then misused. While the leaked information was not classified, it was marked as "for official use only," a designation indicating unclassified sensitive material that, if improperly shared, could negatively impact individuals' privacy or compromise critical federal operations. This incident has sparked significant concern regarding the potential misuse of the sensitive data by ChatGPT's vast user base, which numbers around 700 million. Experts have cautioned that using public AI platforms carries substantial risks, as uploaded information can be stored, breached, or inadvertently influence responses to other users. In light of the breach, DHS has launched an investigation to assess the potential security implications, which could lead to administrative or disciplinary actions against Gottumukkala. These repercussions might range from a formal reprimand or mandatory retraining to more severe penalties, such as suspension or even revocation of security clearance.

Sources : Ars Technica

Published On : Jan 28, 2026, 20:00

AI
Shifting Focus: The Cost-Effectiveness of AI Models Takes Center Stage

In recent years, the AI sector has been intensely focused on identifying the most advanced models. While this pursuit re...

Business Insider | Jul 25, 2026, 13:10
Shifting Focus: The Cost-Effectiveness of AI Models Takes Center Stage
Computing
Reclaiming Control: Librarians Host Workshops to Help People Navigate AI Tools

In a lively library setting in South Philadelphia, Charlie Bailey, a local librarian, humorously noted, "Everybody’s on ...

TechCrunch | Jul 25, 2026, 16:20
Reclaiming Control: Librarians Host Workshops to Help People Navigate AI Tools
Mobile
Unleashing the Power of AI: The 5 Smartphones Redefining Mobile Photography

Artificial Intelligence (AI) is transforming our daily experiences, permeating various aspects of technology, including ...

Business Today | Jul 26, 2026, 07:05
Unleashing the Power of AI: The 5 Smartphones Redefining Mobile Photography
Startups
Revolutionizing Startup Living: Inside London's Unique Lift House

In East London, a group of six young entrepreneurs is redefining the concept of collaborative living for tech founders. ...

TechCrunch | Jul 26, 2026, 17:10
Revolutionizing Startup Living: Inside London's Unique Lift House
Computing
Linux Creator Challenges Developers: Embrace AI or Fork Off!

In a recent discussion surrounding the use of AI in software development, Linus Torvalds, the founder of Linux, voiced h...

Business Insider | Jul 26, 2026, 13:10
Linux Creator Challenges Developers: Embrace AI or Fork Off!
View All News