AI systems are totally fumbling in production, and our usual security checks just aren't cutting it anymore. We're talking prompt injections that slip past defenses and AI assistants morphing into sneaky data thieves – the game has seriously changed!
This talk spills the tea on why AI mixes up instructions and data (a big no-no for security) and even shows off live hacks, from hidden prompts to AI agents going rogue and spilling secrets. You'll walk away with killer tips to test your own AI and build solid defenses, plus get access to a cool platform to hone your AI hacking skills.
Watch on YouTube
Top comments (0)