I recently participated in a CTF with GitHub Copilot as my AI assistant. In this article, I share how AI helped me solve challenges faster, where it reached its limits, and practical tips for anyone interested in using AI in cybersecurity competitions. While I focus on Copilot, the lessons apply to any AI assistant. I’d love to hear about your experiences or opinions on using AI for CTFs or security research!

Kind of weird to hear people talking about this. One of the last classes I took in college (ca 2018 or so) one of my professors said this was going to happen. I remember thinking "oh yeah right, that's some ridiculous sci-fi crap, LSTMs and markov chains can barely generate meaningful English." And now here we are.