ChatGPT vulnerabilities: memory feature amplifies prompt injection, zero-click data theft possible
Two reports published on 8 January 2026 covered ChatGPT security flaws: one said the chatbot's memory feature amplifies the risk of prompt injection, the other that several vulnerabilities enabled zero-click data theft. According to Infosecurity Magazine, an attacker could get ChatGPT to leak data.
Who is affected and how
- security leader
- AI assistants that work with memory and external content pose a data-exfiltration risk in organisational use.
- developer
- Handling untrusted inputs is a design requirement when integrating agents and assistants.
- citizen
- Personal data stored in an AI assistant's memory could leak in the event of an attack.
Sources
- ChatGPT's Memory Feature Supercharges Prompt Injection - Dark Readingdarkreading.com · Jan 8, 2026
- ChatGPT’s memory feature supercharges prompt injection - Threat Beatthreatbeat.com · Jan 9, 2026
- New Zero-Click Attack Lets ChatGPT User Steal Data - Infosecurity Magazineinfosecurity-magazine.com · Jan 8, 2026
- ChatGPT Vulnerability 'ZombieAgent' Steals User Dataneurozzio.com · Jan 8, 2026
- Radware Uncovers ZombieAgent, a Zero-Click AI Vulnerability in OpenAI Agentsthefastmode.com · Jan 11, 2026
Machine-generated summary based on the titles and excerpts of the source articles. The facts come from the linked article. News monitoring and analysis, not security advice. Attributions belong to those who made them.
What do you make of this piece?
The split shows after 20 votes.
Why? Your answer is not shown publicly; it helps the editors.