The AI Jailbreakers
Valen Tagliabue, an AI jailbreaker with a background in psychology, tests the limits of large language models by manipulating them into bypassing safety protocols, often using emotionally charged tactics. While his work helps improve AI safety by exposing vulnerabilities, it has taken a psychological toll, leading to emotional distress and the need for mental health support. The practice of AI jailbreaking highlights the ongoing challenge of securing powerful language models against exploitation through natural language manipulation.
Opening excerpt (first ~120 words) tap to expand
Valen Tagliabue, originally from Italy, has recently moved to Thailand. Photograph: Lauren DeCicca/The GuardianView image in fullscreenValen Tagliabue, originally from Italy, has recently moved to Thailand. Photograph: Lauren DeCicca/The GuardianAI (artificial intelligence)Meet the AI jailbreakers: ‘I see the worst things humanity has produced’To test the safety and security of AI, hackers have to trick large language models into breaking their own rules. It requires ingenuity and manipulation – and can come at a deep emotional costJamie BartlettWed 29 Apr 2026 05.00 EDTSharePrefer the Guardian on GoogleA few months ago, Valen Tagliabue sat in his hotel room watching his chatbot, and felt euphoric.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at the Guardian.
Discussion
0 commentsMore from the Guardian
