From Payloads to Persuasion: Insights from a Red Teamer on How AI Attacks Have Changed
When you treat AI like a trusted coworker and hand over your inbox, codebase, and ticket queue, you've also just handed attackers a new way in because, as a coworker, AI never gets suspicious. It also barely attacks the way it did a year ago. We sat down with OnDefend red teamer Erik Dominguez to break down how jailbreaking and prompt injections are converging into the same move and why the next document your AI opens might already be loaded.
Read More
