.png){kind=logo}
Most students know their school has cameras. Fewer know what happens after a camera, a laptop filter, or a safety app generates an alert. Does it just get logged, or does it quietly change how adults see a student? That question is sitting behind Senate Bill 5956, a Washington proposal that would set new limits on AI and surveillance tools in public schools, especially when discipline is involved.
SB 5956 would prohibit school districts, charter schools, and state tribal education compact schools from using an automated decision system as the “sole or determinative basis” for any discipline related decision. It also says schools could not impose big actions like emergency removal, suspension, expulsion, referral to law enforcement, or assignment to an alternative education setting based only on a prediction, score, or classification generated by an automated system, or on data from school surveillance technology, unless there is independent human investigation and context.
The most controversial part is the bill’s approach to predictive “risk scoring.” SB 5956 would block schools from using automated systems as the deciding factor to generate risk scores or similar predictive labels for individual students. It also limits internal “watchlists” of students that are wholly based on an automated decision system.
The bill draws a bright line on facial recognition in schools, too. The Senate bill report says, without exception, districts may not use facial recognition services for ongoing surveillance, real time or near real time identification, or persistent tracking of students.
SB 5956 also restricts biometric data being used to infer sensitive personal traits, like emotional states or mental health conditions, sexual orientation, gender expression, or gender identity, with limited exceptions listed in the bill report.
At the same time, the bill tries to avoid tying schools’ hands completely. It says nothing should be construed to limit existing authority to discipline students under state law, and it allows analyzing nonbiometric behavior information to determine a “level of concern” if it is used to develop supportive interventions.
As of Jan. 27, 2026, SB 5956 is in Senate Rules (second reading) after being sent to Rules on Jan. 23. If it advances, OSPI would update its AI guidance during its regular review cycle, and WSSDA would publish a model district policy by Feb. 1, 2027.
