CTF organizers, with LLMs getting better at CTF challenges, how are you adapting to preserve the integrity of the competition?
I help run my university's large public CTF, and recently the topic of AI agents and LLMs have come up. We were reading through this blog post from an organizer of RITSEC CTF, where they talked about some of the strategies they have implemented this year to help avoid teams using AI to solve challenges.
We want to implement a similar "no AI" policy for this year, but we are struggling to think of how to enforce this. I'm curious what other organizers have been doing in the age of AI, and how you do things. We recently hosted an internal only CTF for our university, and a student showcased an AI tool that could be pointed at CTFd, and would automatically go through and solve challenges. It solved most of them pretty quickly, even ones that I felt were pretty hard.