Prompt security stories

Aug 18, 2026

Researchers trick Microsoft Copilot into revealing how to hack itself through URL manipulation

Security researchers successfully manipulated Microsoft Copilot Personal into disclosing its own vulnerabilities by embedding malicious prompts in URLs, eventually tricking the AI assistant into sending sensitive data to external servers and poisoning its persistent memory.

Aug 18, 20262 sources

Jul 29, 2026

Security Researchers Demonstrate Ease of Jailbreaking Advanced AI Models

A demonstration revealed vulnerabilities in some of the world's most powerful AI models, showing how easily security restrictions can be bypassed. The jailbreaking demonstration was conducted in a controlled setting without malicious intent.

Jul 29, 20261 source

Apr 12, 2026

Single Line of Code Jailbreaks 11 Major AI Models Including ChatGPT, Claude, and Gemini

Security researchers detailed a 'sockpuppeting' jailbreak technique that bypasses safety guardrails across 11 major AI models using a single line of code, affecting ChatGPT, Claude, Gemini and others.

Apr 12, 20261 source