As we train and release powerful AI systems, we create opportunities for bad actors to use these systems as part of offensive cyber attacks. One possible attack vector is automated, personalized phishing, powered by an AI with access to your public-facing information. However, there is little to no research on the effectiveness of such attacks or how one can defend against them.
That’s why Palisade is going to be testing a variety of AI-powered phishing attacks, with the goal of collecting data on how capable current models are at these kinds of attacks.
If you’d like to participate in our research, please fill out this NOT PRODUCTION READY Google Form.