California's K-12 AI guidance: practical, not panicked

California's K-12 AI guidance: practical, not panicked

The California Department of Education has released updated AI guidance for public schools, covering transitional kindergarten through grade 12. It is advisory rather than mandatory, and it sits alongside existing legal requirements. The guidance puts professional learning front and centre, insists AI must support rather than replace human judgment, and pushes schools to scrutinise how tools collect, store, retain, and use student data. Equity, bias, and access — especially for students in underresourced communities — get explicit attention.

Gobble's Take: "Use it, but read the fine print" is not a strategy. It is, however, a start. Source: EdTech Innovation Hub

What a real district AI policy actually says: oversight, privacy, accountability

Tucson Unified School District's AI policy opens with a clear purpose: set regulations for AI use across education, administration, and operations while protecting privacy and supporting equitable learning. It defines reactive, predictive, generative, and compliant AI systems, applies to everyone from students to third-party developers, and is unambiguous on one point — AI must not replace human educators or staff, and every decision involving AI is subject to human review and approval. Systems must also align with fairness, equity, and inclusivity, and be regularly evaluated for bias or discriminatory outcomes.

Gobble's Take: The best district AI policies are not "yes" or "no." They are "yes, and here is who is still accountable." Source: GovBoard TUSD Policy Code IJND

AI didn't wait for school boards, and neither did students

A 2025 practical guide drawing on voices from over 100 teachers and students opens with a blunt observation: generative AI didn't wait for school boards. Students and teachers were using it before policies existed, and that created urgency and confusion in equal measure. The guide is aimed at educators and school leaders, and its message is equally blunt — schools must act, experiment, and iterate, not wait. It points toward questions about student data privacy, teacher knowledge, class-ready policy modes, and whether AI can save teacher time without swallowing the teacher's voice entirely.

Gobble's Take: The tools are already in the building. "Later" is not on the timetable. Source: WINS Solutions

The cheating question has quietly become an attribution question

A teacher held up two student essays at a planning meeting and said: "I can't tell — which one of these a student actually wrote — and I can't prove anything either way." She wasn't angry. The tools had simply outrun her. Generative AI erased the fingerprints that once made plagiarism visible, and the argument is straightforward: no detector is bringing back the assumption that the work in front of a teacher reflects the mind that turned it in. The central question of every assignment in 2026 is now "whose thinking is this?" — and a September 2025 RAND survey suggests schools have no answer ready. Over 80% of students said their teachers had not explicitly taught them how to use AI for schoolwork. 54% said they were using it anyway. Only 45% of principals reported having any AI policy at all.

Gobble's Take: Schools that skip the attribution conversation will spend a lot of time guessing at the answer. Source: Think Forward

In Case You Missed It

Yesterday's top stories:

• Proofademic is the best AI detector for teachers in 2026
• Teachers are catching AI use, but most responses are talk, not punishment
• The real homework question is shifting from “Did you use AI?” to “How do we judge the work?”