Washington has given public bodies another year. It should not become another year of audits, arguments, and avoidable delay.

The rule moved, but the obligation did not

The U.S. Department of Justice has pushed back the compliance dates for its digital accessibility rule under Title II of the Americans with Disabilities Act. The change was made through an Interim Final Rule published in the Federal Register on April 20, 2026. Larger state and local government entities now have until April 26, 2027. Smaller public entities and special district governments now have until April 26, 2028. [1]

That is the news, but it is not the whole story.

The Department did not abandon digital accessibility or say that disabled people can wait. It did not say that inaccessible websites, forms, apps, videos, and documents are acceptable for another year. The underlying 2024 rule still sets WCAG 2.1 Level AA as the technical standard for state and local government web content and mobile apps. It also applies when those services are provided through contractual, licensing, or other arrangements. [2]

That distinction matters because the market will be tempted to describe the change as a pause. The timetable has moved, but the obligation to provide access has not.

The Department’s own first-steps guidance now confirms the extended dates, but still frames the task as one of compliance planning and action. [3] The ADA.gov fact sheet also makes clear that even where an exception applies, state and local governments still have duties around effective communication, reasonable modifications, and equal opportunity. [2]

So the phrase "another year to sort this out" is the wrong reading. A better reading is this: Washington has given public bodies more time because the current delivery model has made accessibility look slower, more expensive, and harder to deliver than it should be.

The cost case exposes a delivery problem for access

The Justice Department’s reasoning deserves close attention. The Department said it had overestimated the staffing and technology capabilities of covered entities. It pointed to resource constraints, staff availability, litigation risk, and the difficulty of relying on technology to make web content and mobile apps accessible within the original dates. [1]

The Department’s cost numbers are just as important to this argument. The Department estimated a first-year implementation cost of $16.949 billion and recurring annual costs of $1.99 billion. It also estimated that delaying implementation by one year would generate $2.775 billion in present-value cost savings over a 10-year period. Small entities were expected to account for $1.472 billion of those savings. [1]

Those numbers do not prove that accessibility is too expensive. Instead, they prove that the current model of delivering accessibility is being understood by government as expensive.

That difference matters because access is not the cost problem; the method is the cost problem.

For years, digital accessibility has often been pursued through audit cycles, manual remediation, expert sign-off, technical qualification, and legal risk management. Much of that work is necessary, and some of it is valuable. But when the entire market defaults to expert-led manual correction, public bodies are left with a predictable conclusion. They believe the work requires more people, more time, and more budget than they have.

That belief helped make regulatory delay easier to defend.

The federal delay is not the only timetable that matters

A market-focused view also needs to avoid a second mistake. The ADA Title II delay is significant, but it is not the only accessibility timetable in operation.
The Department of Health and Human Services has separate Section 504 requirements for web content, mobile apps, and medical kiosks made available by recipients of federal funding from HHS. Those requirements adopt WCAG 2.1 Level AA. Recipients with 15 or more employees must comply by May 11, 2026. Smaller recipients must comply by May 10, 2027. [4]

State obligations also continue under frameworks such as Colorado’s accessibility law. Colorado’s accessibility framework, for example, treats digital access as part of its state disability law framework. It identifies potential penalties, civil action, court-ordered compliance, damages, attorney’s fees, and a statutory fine where covered government entities fail to comply. [5]

This is why the new federal timetable cannot sensibly be treated as permission to wait. Public bodies and publicly funded organisations still face legal, operational, and reputational exposure. The wider direction of travel has not changed. Digital access is becoming more specific, more measurable, and more enforceable.

The expert economy now needs a harder examination

Accessibility campaigners are right to be concerned when deadlines move. Disabled users have already waited long enough for access to public services that many other people take for granted.

But the reaction to the delay should not stop at criticism of Washington. It should also examine the professional economy that has grown around accessibility.

That economy includes consultants, auditors, remediation vendors, platforms, training providers, lawyers, and expert witnesses. This article is not an attack on the legal profession, and it should not be read as criticism of everyone working in accessibility. Many experts have raised standards, forced accountability, and helped organisations improve.

The problem is different because complexity starts to protect itself when expertise becomes the only accepted route to progress.

Some professionals make the problem sound more fragile than it is. Some vendors price the work as if every barrier requires specialist manual intervention. Some platforms build dependency by turning discovery into repeat discovery, and remediation into recurring consultancy. The result is a market where everyone says access matters, while the work keeps looking too expensive to complete.

This is the uncomfortable commercial question raised by the DOJ delay: who benefits when accessibility is made to look unaffordable?

Disabled users do not benefit because they are made to wait longer. Public bodies do not benefit because they carry unresolved exposure. Campaigners do not benefit because their objective moves further away. The group most protected by complexity is the part of the professional market that sells complexity back to the organisations trying to comply.

Automation should change what gets done first in practice

The accessibility profession is right that automation is not perfection. Automated testing misses context, and AI can still misread intent. Human judgement remains essential for service design, legal confidence, content meaning, user testing, and final assurance.

The Justice Department made a related point in the Interim Final Rule. It said advanced technology, including generative AI, does not yet reliably automate remediation of inaccessible content at scale. It also said that if public entities cannot rely on technology, they will need to rely more heavily on manual work instead. [1]

That point should be taken seriously, but it should not become a veto against automation.

The useful distinction is not between humans and AI. The useful distinction is between work that needs judgement and work that does not.

Many barriers are repetitive, mechanical, structural, and known. WebAIM’s 2026 analysis of one million home pages found that 95.9% had detected WCAG failures. It also found average home-page complexity had increased to 1,437 elements, and that users with disabilities would expect to encounter detected errors on one in every 26 page elements. [6]

The same report found that 96% of all detected errors fell into six recurring categories: low contrast text, missing alternative text for images, missing form input labels, empty links, empty buttons, and missing document language. [6]

That matters because it exposes the scale problem and the opportunity at the same time. The web is not being fixed fast enough by manual correction. At the same time, many of the common barriers are precisely the kind of work where AI-led automation can do the majority of the heavy lifting, with human insight applied where judgement is needed.

Research is already moving in that direction with measurable results. A 2025 arXiv paper on AccessGuru proposed combining accessibility testing tools and large language models to detect and correct web accessibility violations in HTML code. The authors reported an average violation score decrease of up to 84% against their benchmark, while also describing automatic correction as an open challenge rather than a solved issue. [8]

That is the right balance for any defensible policy approach. AI should not be sold as instant legal perfection. It should be used as a way to remove many barriers faster, reduce the cost of first improvement, and reserve expert time for the cases where expert judgement genuinely matters.

Legal activity does not always create practical access

Litigation also deserves scrutiny because legal activity and practical access are not the same thing.

Seyfarth Shaw reported that ADA Title III federal lawsuit filings remained high in 2025, with 8,667 filings that year. That figure reflects a continuing enforcement environment, especially in states such as California, Florida, and New York. [7]

Lawsuits can force attention, create pressure, and result in settlements, deadlines, and public statements. But they do not always produce durable access.

Reuters reported in February 2026 that the Justice Department objected to a proposed website accessibility settlement involving Fashion Nova. The report said the claims website itself allegedly had barriers for blind users, and that the proposed lawyers’ fees and costs were greater than the remaining amount available to class members. [7]

That case does not define the entire sector. It does show why the market should not mistake process for progress. Access is not achieved because a settlement exists. Access is achieved when people can actually use the service.

The people who lose are not the vendors

The DOJ delay is therefore not only a legal event; it is a mirror held up to the accessibility market.

If public bodies cannot afford to fix the problem at scale, the question is not whether access matters less. The question is whether the current method is fit for the scale of the problem.

The people who lose from delay are not the vendors. The people who lose are disabled users, residents, students, patients, claimants, parents, employees, and anyone else who depends on public digital services.

The public bodies also lose because a one-year extension does not remove the underlying work. It does not remove operational risk, complaints, litigation exposure, service failures, or reputational damage. It only moves the date by which one defined standard must be met.

The professional market should be honest about this. If the cost of access has become a reason for delay, then the market must ask whether its delivery model is serving access or serving itself.

The extra year should change the method, not lower ambition

The extra year should not become another year of audits, arguments, and expensive hesitation. It should be used to change the way accessibility work is done.

That means using automation to identify and correct high-volume barriers quickly. It means using AI to do the majority of the mechanical work. It means applying human insight where context, legal confidence, usability, and content meaning demand it. It means measuring improvement, not just documenting failure.

That approach is not anti-expert; it is a better use of expertise.

The purpose of accessibility law is not to create work for experts. Its purpose is to make the web available to all.

The DOJ delay should not weaken that purpose, and it should make the policy lesson clearer. If the old model makes progress look unaffordable, then the answer is not to lower ambition. The answer is to change the delivery model.

Washington has given public bodies more time, and the market should not waste that time defending a method that has already failed to scale.

Source links

The source links below are included as plain URLs for publication checking, legal review, and editorial verification.

[1] Federal Register, DOJ Interim Final Rule
Why it matters: Confirms the one-year extension, new compliance dates, cost reasoning, staffing concerns, technology concerns, and the Department’s stated position on generative AI and manual work.
Link: https://www.federalregister.gov/d/2026-07663

[2] ADA.gov, DOJ Title II web rule fact sheet
Why it matters: Confirms WCAG 2.1 Level AA requirements, Title II scope, contractor coverage, exceptions, and continuing ADA obligations.
Link: https://www.ada.gov/resources/2024-03-08-web-rule/

[3] ADA.gov, first steps guidance
Why it matters: Confirms the DOJ’s implementation guidance and updated compliance dates for state and local government entities.
Link: https://www.ada.gov/resources/web-rule-first-steps/

[4] HHS Section 504 digital accessibility requirements
Why it matters: Confirms separate HHS-funded entity deadlines, WCAG 2.1 Level AA requirements, and continuing duties during the interim period.
Link: https://www.hhs.gov/sites/default/files/new-requirements-accessibility-web-content-mobile-apps-kiosks.pdf

[5] Colorado Office of Information Technology
Why it matters: Shows that state-level accessibility duties continue and are not erased by federal delay.
Link: https://oit.colorado.gov/standards-policies-guides/guide-to-accessible-web-services/faq-hb21-1110-colorado-laws-for-persons

[6] WebAIM Million 2026
Why it matters: Provides evidence that web accessibility barriers remain widespread, common errors are concentrated, and page complexity is increasing.
Link: https://webaim.org/projects/million/

[7] Seyfarth Shaw ADA Title III lawsuit filings and Reuters Fashion Nova settlement scrutiny
Why it matters: Supports the article’s point that legal pressure remains active and that legal activity does not always equal practical access.
Link 1: https://www.adatitleiii.com/2026/02/ada-title-iii-federal-lawsuit-filings-fall-slightly-to-8667-in-2025/
Link 2: https://www.reuters.com/legal/government/an-ada-class-action-settlement-draws-doj-scrutiny-irony-2026-02-04/

[8] arXiv, AccessGuru research paper
Why it matters: Supports the cautious argument that AI-assisted accessibility remediation is advancing while still requiring oversight and judgement.
Link: https://arxiv.org/abs/2507.19549