Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Data Security topic
No spam. Unsubscribe anytime.
District says AI-detection tools aren’t yet reliable; emphasizes teacher monitoring and data security
Summary
Board heard tests of AI-detection tools with roughly 75% accuracy in local checks and a recommendation against purchasing detectors now; staff said data-security training and auditing of user access remain priorities.
During Q&A after the AI demonstration, a community member asked whether the district was using AI-detection tools to identify machine-generated student work. Laurie Gregory and other presenters said regional curriculum groups and the tech consortium had considered detectors, but local testing was mixed. "There's usually, like, maybe a 75% accuracy rate," Gregory said of the detectors she tested, and she described instances where generated text was identified as human by detectors. Staff concluded current detectors are not sufficiently reliable to warrant purchase districtwide and stressed teacher strategies — drafting in class, peer review and monitoring through tools such as GoGuardian — as the primary safeguards.
Gregory also summarized district data-security practices: regular staff training, auditing program access and documenting data flows to limit unauthorized access. She said the district will continue to attend trainings and refine its approach to emerging risks rather than rely solely on automated detectors.
