Reading the April 2026 IFR: What the DOJ Said About AI-Assisted Compliance (and What It Didn't)
The DOJ's April 2026 IFR extended the ADA Title II deadlines but tightened the standard for what counts as compliance. A close reading for university counsel.
What the IFR actually says
For general counsel briefing university leadership on the April 2026 Interim Final Rule (RIN 1190-AA82), most secondary coverage in the days after publication led with the date extension and stopped there. That coverage is incomplete in a way that matters for board reporting. The substantive language in the IFR's preamble is more consequential than the date change itself, because it constrains what an institution can credibly claim as a compliance strategy in any complaint investigation filed from April 20, 2026 forward — and that constraint applies whether the complaint is filed before or after the new dates of April 26, 2027 (large public entities) or April 26, 2028 (smaller entities and special districts). For the broader defensibility framework that the rest of this article maps onto, see the defensibility standard. This article narrows in on the IFR text itself.
The mechanics of the IFR are narrow. It revises 28 CFR Part 35 to extend two compliance dates. It does not amend the technical standard — WCAG 2.1 Level AA remains the conformance benchmark. It does not amend the scope — every covered entity remains covered, every category of web content and mobile app remains in scope. It does not create new defenses, exemptions, or safe harbors. The substantive obligations of the 2024 Final Rule (28 CFR Part 35, Subpart H) — non-discrimination, equal access, ongoing remediation — were in effect before April 20, 2026 and continue without modification.
What the IFR adds is reasoning. The DOJ's preamble explains why the original two-year compliance window proved insufficient, and the reasons named in that explanation will be cited back to institutions in any future investigation. The single most consequential sentence in the preamble — for procurement, general counsel, and any technology vendor selling into the higher-ed accessibility market — is the regulator's own statement that "advanced technology, such as generative AI, does not yet reliably automate the remediation of inaccessible content at scale, and staff resources and availability continue to pose significant challenges." A second passage, summarising STEM-content concerns, records that "generative AI...cannot reliably automate the remediation of STEM materials at scale, and human oversight is required to ensure accessibility."
These sentences are now part of the federal record. They have the same evidentiary status as any other regulatory finding. An institution arguing in a future investigation that an AI-only tool produced compliance is arguing against the regulator's own published characterisation of what AI-only tooling can do. That is a strategic problem the IFR's preamble created on April 20, 2026, regardless of when an institution's compliance date falls.
What the IFR doesn't say
Equally important is what the IFR does not contain. Five categories of exemption that a casual reader might infer from the extension are absent from the text.
There is no AI-tool safe harbor. The IFR does not authorise reliance on automated remediation, with or without disclaimers. The opposite is true: the preamble's discussion of generative AI's limitations is the federal record's first sustained statement of regulatory skepticism about AI-only approaches to web accessibility.
There is no low-traffic exemption. Coverage extends to every public-facing web page and every document linked from those pages, irrespective of how many users access them. The volume of an institution's archive is not an excuse — it is a reason for a strategy that scales.
There is no resource-constraint exemption. Resource constraints are cited in the preamble as a reason for additional time, not for non-compliance. An institution that arrives at the new compliance date without a documented programme cannot retroactively claim resource constraints as a defence — the IFR already acknowledged the constraint when it granted the extension.
There is no automatic deadline reset for content published after the rule. Pages, documents, and applications added between April 20, 2026 and the new compliance date are subject to the same conformance standard as legacy content. An institution that uses the additional year for new initiatives without applying the standard to that new content is creating fresh exposure.
There is no relief from ongoing obligations. The duty to remediate identified barriers, to respond to complaints, and to maintain non-discriminatory programmes is a continuous one. The compliance date is the latest point at which the conformance baseline must be reached; it is not the only date on which the institution can be evaluated. A complaint filed on May 15, 2026 is evaluated against the same WCAG 2.1 AA standard that applied on April 23, 2026. The investigation timeline does not pause for the IFR.
The picture that emerges from these absences is a regulator that extended the timeline because the original deadline was unachievable, not one that softened the substantive expectation. For general counsel, that is the operative posture to report up the chain.
What general counsel should brief the board
A board briefing on the IFR should resolve five points. Each maps to a decision the board can take in the current fiscal-quarter cadence rather than waiting for the new compliance date.
First, the deadline moved but the standard tightened in the regulatory record. The IFR's preamble is now a document that any complainant or investigator will cite. Programme strategy should be evaluated against the substance of that preamble, not its calendar.
Second, AI-only tooling is on the federal record as inadequate. Any vendor or internal team representing AI-only remediation as sufficient is taking a position the regulator has already characterised as unreliable. Procurement decisions made before April 2026 may need re-evaluation in light of this characterisation.
Third, the institution's existing programme should be audited against the question the IFR's preamble actually asks: is there a documented process producing remediated content at a scale that closes the gap before the new date? "Documented" here is the load-bearing word — see the defensibility standard for what documented process means in OCR-investigation language.
Fourth, procurement evaluation criteria should reset around process and documentation rather than scanner-score uplift. The contract clauses that survive an investigation are clauses about audit trails, human-review SLAs, and inference-location disclosures — not about percentage improvements on automated scans.
Fifth, the new dates are a floor, not a ceiling. Institutions that achieve substantial conformance ahead of the April 2027 or April 2028 dates and can demonstrate documented progress are in a meaningfully different defensibility posture than institutions that arrive at the deadline date claiming they used the extension well. The IFR did not move the standard. It moved the worst possible date by one year — and a programme that finishes on that date with no margin is a programme that ran the clock to zero.
The IFR's window of additional time is most usefully read as exactly that: a window for better compliance, not for delayed compliance. The regulatory record is now framed in a way that makes the difference between those two readings consequential to any future investigation. General counsel and accessibility-programme leadership should brief their boards accordingly.

RD (Reg) Crampton
•Founder & CEOFounder, CEO & lead developer of Aelira. Passionate about making education accessible to everyone. Building the tools universities need to meet accessibility compliance.
Related Articles
Why Compliance Scores Aren't Defensible: The Case for Human-Review Audit Trails
Scanner scores measure conformance under ideal conditions. What regulators actually want to see is a per-document record of human review. Here is what that record needs to contain.
The 2027/2028 Math: Why Manual-Only Remediation Cannot Meet the Extended Deadline
The April 2026 IFR gave universities an extra year. The math says it does not help — manual-only remediation cannot meet 2027/2028 at any institution with a serious archive.
The Defensibility Standard: Why Human-in-the-Loop AI Is the Only Compliant Model for University Accessibility Programs
After the April 2026 IFR, university accessibility compliance is judged on documented process, not scanner scores. Here is what that means for procurement.
Ready to achieve accessibility compliance?
Join the pilot program for early access to Aelira's AI-powered accessibility platform
Apply for Pilot