AI Training Alone Didn't Work. LAUSD Proved It
LAUSD required digital-citizenship training before AI access, then suspended it anyway. Training solves literacy, not evidence.
Anshika Verma · Founder, GrowWise · Last updated
Before September 2026, LA Unified allowed students 13+ to use approved AI tools only after completing digital-citizenship instruction and reviewing a responsible-use policy - the training-first model widely recommended as the responsible middle ground. The district suspended it anyway for the 2026–27 year. The lesson: AI literacy training addresses a real gap, but not the one that determines whether a school can actually trust student work.
LA Unified Required AI Training Before Student Access. It Still Wasn't Enough to Keep the Policy Alive.
If your school's current AI plan is "we'll train students on responsible use, then allow access," it's worth knowing that the second-largest school district in the country ran exactly that plan and pulled it back.
LAUSD's prior policy required students 13 and older to complete digital-citizenship instruction and review a responsible-use policy before touching any approved AI tool. That is the standard advice: don't ban outright, educate first, then permit supervised use. Consultants, professional-development vendors, and a good chunk of published AI-in-education guidance recommend some version of this exact sequence.
It didn't survive. For the 2026–27 school year, LAUSD suspended student-facing generative AI district-wide while a Generative AI Ad Hoc Committee reviews policy from the ground up.
What training actually solves
Digital-citizenship instruction is genuinely useful for what it's built for: helping a student understand what AI is, where it can be wrong, how to cite it, and what responsible use looks like conceptually. That's a real gap, and closing it matters.
What training can't solve
Training doesn't give a teacher any new information about a specific student's specific work. A student can complete a thorough digital-citizenship module and still turn in an AI-generated essay the following week - the training changes what the student knows about responsible use, not what a teacher can verify about any individual piece of work. Those are separate problems, and only one of them gets solved by a module.
This is likely close to the actual gap that made LAUSD's policy hard to sustain: a training requirement addresses literacy, but a school still needs some way to know, assignment by assignment, whether a particular student's understanding is real. Training alone doesn't produce that evidence. It never did.
The distinction that actually matters going forward
There's a version of "AI literacy" that's necessary and a version of "AI access control" that's a different problem entirely, and conflating them is likely part of why the training-then-access model hasn't held up under real deployment at scale. Keep the literacy piece - students should understand what AI is and its limits, and NYC's own moratorium explicitly preserves AI literacy instruction even while restricting hands-on use.
But the access-control piece needs a different kind of solution: not a training gate a student passes once, but a live diagnostic layer that shows a teacher what a specific student actually understands, continuously, regardless of what technology they used to produce any single piece of work. That's a structurally different tool than a training module, solving a structurally different problem - and it's the piece LAUSD's otherwise well-designed policy didn't have.
What this means for schools building a policy now
Training is worth keeping. It just can't be the whole plan. Pair AI literacy instruction (what LAUSD had) with a way to verify actual student understanding independent of any single assignment (what LAUSD didn't have) - evidence from in-class assessment, diagnosed at the level of the specific skill, checked by the teacher, not assumed because a module was completed months earlier.
See Teacher-Facing AI in Action.
Related reading
FAQ
What was LAUSD's AI policy before the September 2026 ban?
Students 13 and older could use school-approved AI tools after completing digital-citizenship instruction and reviewing a responsible-use policy.
Why wasn't training-then-access enough to sustain the policy?
Training addresses whether a student understands responsible AI use conceptually, not whether a teacher can verify what a specific student actually understands on any given assignment. Those are different problems.
Does this mean AI literacy training is a waste of time?
No. NYC's own moratorium explicitly preserves AI literacy instruction alongside its restriction on hands-on student use - literacy and access control are separate, both-necessary pieces.
What should schools pair with AI literacy training?
A way to verify actual student understanding independent of any single assignment - in-class assessment diagnosed at the level of specific skills, checked continuously by the teacher.
Sources
Frequently asked questions
- What was LAUSD's AI policy before the September 2026 ban?
- Students 13 and older could use school-approved AI tools after completing digital-citizenship instruction and reviewing a responsible-use policy.
- Why wasn't training-then-access enough to sustain the policy?
- Training addresses whether a student understands responsible AI use conceptually, not whether a teacher can verify what a specific student actually understands on any given assignment. Those are different problems.
- Does this mean AI literacy training is a waste of time?
- No. NYC's own moratorium explicitly preserves AI literacy instruction alongside its restriction on hands-on student use - literacy and access control are separate, both-necessary pieces.
- What should schools pair with AI literacy training?
- A way to verify actual student understanding independent of any single assignment - in-class assessment diagnosed at the level of specific skills, checked continuously by the teacher.
See teacher-facing AI in action
AI that helps teachers see misconceptions and decide what to reteach next, without student chatbots.
2026 AI-in-schools policy comparison · Five questions before signing an AI vendor