LOWVulnerability
Global

OpenAI Patches ChatGPT Data Exfiltration Flaw and Codex GitHub Token Vulnerability

Monday, March 30, 2026 at 06:05 PM UTC·Source: The Hacker News

Updated: Wednesday, April 1, 2026 at 07:13 PM UTC

Executive Summary

A previously unknown vulnerability in OpenAI ChatGPT allowed sensitive conversation data to be exfiltrated without user knowledge or consent, according to new findings from Check Point. "A single malicious prompt could turn an otherwise ordinary conversation into a covert exfiltration channel, leaking user messages, uploaded files, and other sensitive content," the cybersecurity company said in

Analysis

A previously unknown vulnerability in OpenAI ChatGPT allowed sensitive conversation data to be exfiltrated without user knowledge or consent, according to new findings from Check Point. "A single malicious prompt could turn an otherwise ordinary conversation into a covert exfiltration channel, leaking user messages, uploaded files, and other sensitive content," the cybersecurity company said in
Source Attribution

Originally published by The Hacker News on Mar 30, 2026.

Related Threats

MEDIUMVulnerability

Geopolitics, AI, and Cybersecurity: Insights From RSAC 2026

AI-driven threats, global leadership shifts, and the future of cybersecurity in a rapidly evolving landscape were among the discussions at RSAC 2026 Conference.

Dark Reading
MEDIUMVulnerability

House Dems decry confirmed ICE usage of Paragon spyware

The trio of Democrats weren’t satisfied with Immigration and Customs Enforcement answers, and criticized the spyware’s use. The post House Dems decry confirmed ICE usage of Paragon spyware appeared first on CyberScoop .

CyberScoop
MEDIUMVulnerability

Not Toying Around: Hasbro Attack May Take 'Weeks' to Remediate

The company's 8-K filing notes "unauthorized access" and that it's activated business continuity plans and taken some systems offline.

Dark Reading