San Bernardino City Unified outlines cautious AI rollout for classrooms, highlights district-vetted tools and guardrails

San Bernardino City Unified School District officials on May 19 told the board they are pursuing a cautious, phased rollout of artificial intelligence tools in classrooms that centers human teachers and builds AI literacy for students and staff.

At a presentation led by Ms. Applegate from Educational Services and Dr. Cleveland of the AET team, the district sketched a three‑part approach: (1) build AI literacy, (2) provide professional learning and coaching for educators, and (3) pilot and scale vetted tools that protect student data and align to district curriculum. "We approach it as a tool, not the goal," Dr. Cleveland said, adding the district is focused on "ethical use, transparency, collaboration and values." Ms. Applegate emphasized the district’s intention to keep humans at the center: "Technology is amazing, but it can never replace relationships."

Why it matters: The board’s discussion shows a large district wrestling with how to harness emerging tools while guarding against misuse and privacy risks. Presenters said the initiative aims to help students become "AI literate" for future careers and to free instructional time through targeted tools.

What the district showed and claimed: District slides and presenters named the primary tools they are scaling: Brisk (a writing feedback tool), Formative (real-time assessment), NotebookLM and Google Gemini, as well as district use of OpenAI's ChatGPT via EDU platforms. Presenters cited a 449% uptick in ChatGPT/EDU usage on district-tracked platforms and an estimated 36,420 hours saved (figures reported from the district’s analytics slide). Dr. Cleveland highlighted Brisk’s classroom impact, saying the tool has delivered about "37,000 pieces of individualized feedback" and has helped reduce zero scores on writing prompts (the district cited an 18.7% reduction over a two-year period).

Board concerns and discussion: Board member Mr. Tillman pressed administrators on limits of district control, telling the room: "We do not control AI. We have no way of knowing what AI is going to tell their student." He urged that parent-facing communications include a clear disclaimer that the district cannot guarantee or fully control AI outputs students may encounter outside district-approved tools. Administrators acknowledged the concern and said the district cannot prevent students from using personal devices but that district‑issued devices and selected classroom tools are vetted and monitored. IT staff noted district devices block VPNs and the district uses monitoring tools (cited: Gaggle) to flag concerning activity.

On 'district‑approved tools': Several trustees asked whether policy language should require only district‑approved tools. Administrators recommended retaining a stance that classroom use should be "district-approved" while clarifying that personal devices remain beyond district control; President Ceballos suggested removing the word "only" from presentation language to avoid implying total control.

Next steps: The district plans summer professional learning aimed at embedding vetted AI tools into ELA and math instruction and continued work with the county and CDE to ensure age-appropriate implementation. Presenters said they will provide additional detail to the board on external‑monitoring tools and parent resources.

Ending: The presentation closed with classroom examples and a short video demonstration; presenters reiterated that implementation will be iterative and tied to curriculum goals rather than tool adoption for its own sake.