Episode 184 - AI prompt injection is a security nightmare
Added 2024-03-15 11:54:00 +0000 UTCPrompt injection attacks are a major new security vulnerability that has no clear solution in sight. It's similar to JavaScript vulnerabilities, which are one of the most frequently exploited attack vectors by hackers. All chatbots and generative AI models are vulnerable to prompt injection attacks.
https://www.wired.com/story/chatgpt-prompt-injection-attack-security/
https://www.wired.com/story/generative-ai-prompt-injection-hacking/
