Last updated:
Estimated reading time: 6 minutes
On August 20, 2026, the U.S. Department of Education issued a Dear Colleague Letter on the responsible use of technology in the classroom. The new education technology guidance sets no screen time cap and no minute limits. It does something districts will feel much faster than a cap: it tells states and districts to judge every classroom tool by whether it demonstrably improves learning, and to carry that test into procurement. For anyone who signs an ed-tech contract this fall, the letter reads less like a philosophy statement and more like a purchasing checklist.
Table of contents
What the Department actually released
The document is a Dear Colleague Letter from the Office of Elementary and Secondary Education, announced the same day in a Department press release. The verified facts:
- Date: August 20, 2026.
- Signed by: Kirsten Baesler, Assistant Secretary for Elementary and Secondary Education.
- Status: non-regulatory guidance. It creates no new requirement and preempts no state law.
- Structure: six short sections, from “Holding Educational Value Central” through recommendations addressed separately to states and to education technology providers.
- Lineage: it follows the Department’s July 22, 2025 Dear Colleague Letter on artificial intelligence, signed by Secretary Linda McMahon.
The sentence that organizes the whole letter is this one: “The question should not be whether teachers and students use technology. The question should be whether the technology we ask teachers and students to use improves learning and students’ academic outcomes.”
The five questions every classroom tool must answer
The letter’s operational core is a set of five questions the Department says all education technology should be able to answer. Treated as a purchasing grid rather than a talking point, they are the most useful thing in the document.
| The Department’s question | What a district should require as proof |
|---|---|
| What learning problem does it solve? | A named instructional bottleneck, not a feature list |
| When should it be used? | Placement in the schedule: core block, intervention, station, homework |
| For whom should it be used? | The student population and the tier, with exclusions stated |
| For how long should it be used? | A recommended dose per week, set by the vendor, not improvised by staff |
| What evidence demonstrates improved learning? | Third-party study, local outcome data, or an agreed pilot design |
Most vendor one-pagers answer question one and skip questions four and five. That is precisely the gap the letter is pointing at.
What the letter does not say
It does not set a screen time limit, and that omission is deliberate. Baesler told Education Week that “the public policy debate should not be framed as a choice between technology and no technology,” and the letter asks states to evaluate tools “based on demonstrated learning outcomes and instructional value rather than screen time alone.”
That places the guidance in tension with a real policy wave. According to Education Week and K-12 Dive, more than a dozen states introduced screen-time legislation in 2026 and a handful passed it, Iowa capped elementary digital instruction at 60 minutes a school day, and Los Angeles Unified and Hillsborough County in Florida have adopted district-level caps. The Department is not overriding any of that. It is arguing that a minute count is the wrong instrument, because the same device that delivers low-value scrolling also delivers the assistive technology the letter singles out: “text-to-speech, speech-to-text, real-time captioning, and visual supports.”
The practical read for a district in a capped state: the cap still binds, and this guidance is the argument you use to decide which minutes to cut.
Why this matters for procurement
The letter anchors itself in the evidence language of the Elementary and Secondary Education Act as amended by ESSA, and it is unusually specific about what counts: “External evidence from a third party, randomized controlled trials, implementation reviews, educator feedback, parent feedback, and local outcome data each contribute to informed decision-making and continuous improvement.”
Three things change in the buying cycle when a district takes that seriously.
- Evidence moves into the solicitation. An effectiveness question stops being a scoring nicety at the end of an RFP and becomes a threshold requirement. This matters most on the paths that skip a competitive bid: a sole source justification or a purchase through a purchasing cooperative now needs the same evidence file that an RFP would have forced.
- Renewal becomes a decision again. The letter asks districts to “establish review processes monitoring implementation and outcomes.” Auto-renewal is the default in most K-12 software stacks, and it is exactly what a review process is meant to interrupt.
- Removal becomes a named action. Among the Department’s recommendations is removing tools that fail to improve learning outcomes. That is the sharpest sentence in the letter, and the one vendors will feel.
The arithmetic is manageable, which is why it is hard to argue against. A district running 40 licensed digital tools, at two hours of review each, spends 80 hours a year on the whole portfolio. That is roughly two weeks of one coordinator’s time to defend every renewal in the stack. For more on how these paths differ, see our guide to how K-12 districts buy ed-tech.
Which tools are exposed, and which are not
Read as a survival test, the letter sorts the average district software stack quickly.
| Category | Position under the guidance |
|---|---|
| Assistive technology | Explicitly protected: access to grade-level instruction |
| Diagnostic, formative and summative assessment | Protected: feeds instructional adjustment |
| Practice tools with third-party outcome evidence | Defensible on the letter’s own terms |
| Engagement and gamification with no outcome data | Exposed: high usage, unproven learning |
| Overlapping licenses nobody has reviewed | Exposed: the first line to cut in a review cycle |
Note what this means for language programs. The letter never mentions English Learners or Title III, but EL tools sit on both sides of that table. Translation and vocabulary apps are usage-heavy and evidence-light, while anything that produces measurable oral proficiency growth is on the defensible side. Our review of apps for English language learners sorts that category the same way.
Questions district leaders should ask this fall
- How many digital tools are we licensed for, and who owns the list?
- For how many of them can we answer all five of the Department’s questions today?
- Which contracts auto-renew before we next meet, and what stops that renewal?
- What local outcome data do we already collect that could serve as evidence?
- If our state has a screen time cap, which minutes are we cutting, and on what basis?
- Does our sole source or cooperative purchase file contain an evidence section at all?
Frequently asked questions
Is the August 2026 education technology guidance binding?
No. It is a Dear Colleague Letter, which is non-regulatory. It creates no new legal obligation and does not preempt state screen time laws. Its weight comes from signalling how the Department reads the evidence provisions of ESSA.
Does the guidance set a screen time limit?
No. It sets no numeric limit and no age or grade bands. It asks states and districts to distinguish recreational from instructional use and to evaluate tools on demonstrated learning outcomes rather than duration.
What evidence does the Department expect districts to collect?
The letter names third-party external evidence, randomized controlled trials, implementation reviews, educator and parent feedback, and local outcome data. Randomized controlled trials sit at the top of that list, but local outcome data counts, which puts the standard within reach of a normal district.
How does this relate to the 2025 AI guidance?
The July 2025 Dear Colleague Letter told districts what federal grant funds could support in artificial intelligence. The August 2026 letter is broader and sets the test any tool, AI or not, has to pass. Districts evaluating AI purchases should read them together with our overview of AI for education.
The takeaway
Districts spent five years being told that screen time is the problem. The Department has just told them the measurable thing is learning, not minutes, and has handed them five questions with which to say no. The tools that survive the next procurement cycle will be the ones whose vendors can answer question five without changing the subject.
Related reading
- How K-12 Districts Buy EdTech: Cooperative Purchasing, Sole Source and RFPs
- Sole Source Justification for EdTech: State Rules, the Four Tests, and a Template
- Complete Guide to English Learners in U.S. Schools
- AI in Education: Guides for District Leaders
Sources: U.S. Department of Education press release, August 20, 2026; Dear Colleague Letter on education technology and screen time (PDF); Dear Colleague Letter on artificial intelligence, July 22, 2025 (PDF); Education Week; K-12 Dive.
What a District Can Actually Prove in May
Renewal season has its own quiet moment. A coordinator opens the license list, and for most lines on it the honest answer to “did this help our students” is that nobody knows. The usage report shows logins, sessions and minutes on task. None of those is learning, and everyone in the room understands the difference.
That gap is not a failure of attention. Most classroom software was never built to produce evidence about individual students, so a district ends up being asked to prove something the product never measured. The people who would have to assemble that proof are the same people already running the programme.
Some districts are starting to close the gap at the front end, by asking what a tool will measure before they license it rather than after. Telo AI sits in a narrow group where the practice and the measurement are the same event: each five-minute session updates a WIDA-aligned view of participation and speaking growth, without retaining the conversation itself.
See what a district can show for five minutes of speaking practice
The tools that survive a review are rarely the ones with the best study behind them. They are the ones that were already telling a teacher something useful in November.
Found this guide useful?
Preferred sources show up more often in your Google results, including AI Overviews.