Tomorrow is Global Accessibility Awareness Day. If you are expecting a post full of colored banners, virtual events, and “let’s raise awareness!” energy, keep looking, this is not that post.
Disability advocates across the US are exhausted, and we have more than enough reasons to be. GAAD exists to raise awareness. The goal, always, was to make GAAD unnecessary. We were supposed to be working toward a world where accessibility is so embedded in how we build, buy, and govern that we no longer need a special day to remind anyone that disabled people exist. We are nowhere near that finish line. And in several important areas, we have lost ground.
The celebrations this year are quieter. Some have not happened at all. That is not a coincidence; it is a signal.
AI-Generated Code Is Making the Web Less Accessible, Not More
The accessibility community was cautiously hopeful when AI coding assistants became mainstream. The promise was real: developers who had never considered accessibility might finally receive nudges in the right direction if the tools they used surfaced accessible patterns by default.
That is not what happened.
Instead, developers are now shipping AI-generated code at unprecedented velocity, without the accessibility review that was already insufficient before these tools existed. GitHub Copilot, ChatGPT, Claude, Gemini, and their competitors produce markup that routinely lacks meaningful alternative text, omits or misuses ARIA roles, generates inaccessible form patterns, and fails basic keyboard navigation. The tools are trained on the existing web, which has an accessibility failure rate of almost 96 percent according to the WebAIM Million annual study. You cannot train a model to output accessible code when the overwhelming majority of training data is inaccessible.
We have not solved the accessibility skills gap. Generative AI gives developers a faster way to propagate the same accessibility mistakes at scale. The number of accessibility defects introduced into production is increasing, not decreasing, and the rate at which they are introduced is accelerating. Every month that AI coding tools ship without validated, accessible-by-default output patterns is a month in which the gap between where we are and where we need to be widens.
If you want to follow this intersection closely, the Accessibility and Gen AI podcast, hosted by GAAD co-founder Joe Devon and ServiceNow’s Eamon McErlean, is tracking these developments in real time and is worth adding to your rotation.
Another Year Without a Legally Declared Standard
On April 24, 2025, the Department of Justice published a final rule under Title II of the Americans with Disabilities Act, establishing WCAG 2.1 Level AA as the enforceable technical standard for state and local government websites and mobile applications. That rule marked decades of advocacy. It set clear compliance deadlines, provided government entities with a defined target, and gave people with disabilities a legal basis to demand equal access to the services funded by their tax dollars.
The rule is now facing legal challenge, and its implementation timeline is in jeopardy. The Department of Health and Human Services Section 1557 rule, which imposes similar digital accessibility requirements on recipients of federal health care funding, is also delayed. Both rules have been targeted in the broader federal regulatory rollback that began in early 2025.
Practically, this means that people with disabilities who cannot access a city’s online permit system, a state health agency’s benefit portal, or a public transit authority’s trip planner have no clear federal enforcement mechanism while these challenges work their way through the courts. The discrimination does not pause while lawyers argue. It continues every day. Another year of delay is another year of unequal access for the people these rules were designed to protect. And it bears repeating: Title II only sets the standard for measuring accessibility. The obligation to be accessible already exists.
The Section 504 Fight Is Not Over
In September 2024, Texas Attorney General Ken Paxton filed suit against HHS, joined by 16 other Republican-led states, in a case now known as Texas v. Kennedy. Nominally triggered by the Biden administration’s inclusion of gender dysphoria in the first Section 504 regulatory update in 50 years, the original complaint went significantly further, asking the court to declare Section 504 unconstitutional in its entirety. The regulations under attack also cover accessible medical equipment, the right to receive services in the community rather than institutions, and prohibitions on medical discrimination.
Under sustained public pressure, the states withdrew the direct constitutional challenge in April 2025. However, the case remains active in the Northern District of Texas. The community services provisions, accessible medical equipment requirements, and anti-discrimination framework are still being litigated. Seventeen state attorneys general examined the bedrock of American disability civil rights law and concluded it was a viable litigation target.
Medicaid Cuts and SNAP Work Requirements Are an Attack on Independent Living
The budget reconciliation legislation moving through Congress includes hundreds of billions of dollars in Medicaid cuts and a significant expansion of SNAP work requirements. “Medicaid cuts” may sound like a health insurance argument. For people with disabilities, it is far more than that. Medicaid funds Home and Community-Based Services waiver programs that cover personal care attendants, supported employment, adult day services, and other supports that enable people with disabilities to live in the community rather than in institutions. When these programs are cut, people lose the support that makes independent living possible. Some are forced into institutional care. The consequences are immediate and severe, and they fall disproportionately on people who already have the fewest options.
The expanded SNAP work requirements compound the problem. The episodic nature of many disabilities means a person may be able to work some weeks and not others. The administrative burden of documenting disability status to obtain an exemption is a significant barrier, particularly for people not yet enrolled in SSI or SSDI. Framing these cuts as fiscal responsibility erases who they actually affect. It treats people with disabilities as a budget line rather than as people with constitutionally protected civil rights. That framing should be rejected every time it appears.
The Mental Health Crisis Is a Disability Issue
Mental health conditions are disabilities. That statement should not be controversial in 2026, yet the systems we have built continue to treat mental health as a secondary concern, addressed after everything else is funded, staffed, and prioritized.
Telehealth was a rare expansion of access that genuinely helped people with mental health disabilities. Geographic barriers, transportation barriers, and the energy cost of leaving home during a depressive or anxious episode made in-person therapy inaccessible for enormous numbers of people. Telehealth did not solve every problem, but for many patients it meant the difference between receiving care and receiving nothing. The telehealth flexibilities extended during the COVID-19 public health emergency are now under sustained pressure, with coverage restrictions tightening and reimbursement parity under threat. Pulling back on telehealth access is a direct cut to mental health care for people with disabilities, regardless of how it is framed.
The Mental Health Parity and Addiction Equity Act has been federal law since 2008. Enforcement has been weak for most of that time. Insurers continue to impose tighter prior authorization requirements, lower reimbursement rates, and narrower network standards for mental health services than for comparable medical and surgical benefits. A 2024 report from the U.S. Department of Labor found ongoing, widespread noncompliance. The law exists. The compliance does not.
There is also a dimension of this issue that does not get named directly: the disability rights community is experiencing a mental health crisis of its own. Advocates, many of whom are themselves disabled, have spent years defending rights that were supposed to be settled, fighting rollbacks on every front simultaneously, and watching peers burn out and leave the field. Grief, secondary trauma, and moral injury are not abstractions in this community; they are occupational realities. If we want the advocacy infrastructure to survive long enough to finish this work, we have to be honest about that cost and start treating it accordingly.
Some Improvements May Not Happen in Our Lifetimes
I want to be direct about something that is not said enough: not every accessibility promise will be kept. Not in time. Possibly not ever, for the people who need them now.
Consider accessible subway stations in New York City. The MTA has operated under federal pressure and legal settlements for years. As of 2026, fewer than 30 percent of New York City subway stations are fully accessible. The MTA’s own capital plans acknowledge that full accessibility is not scheduled to be complete until 2055, under current funding trajectories. For wheelchair users and people with mobility disabilities who live and work in New York City, the subway is not a transportation option. It has not been for most of their lives. For many of them, it will not be before they die.
Take flying in your own power wheelchair. The Air Carrier Access Act has been on the books since 1986. In 2024, the Department of Transportation issued a final rule requiring airlines to develop in-cabin wheelchair securement systems on new aircraft by 2035. It is a meaningful step, and it came after decades of advocacy by organizations including All Wheels Up. But the timeline means that people who need this accommodation right now will spend years, possibly decades more, facing a binary choice: transfer to an airline seat that may not safely accommodate their body, or simply not fly. For many people with disabilities, the ability to travel is not a luxury. It is access to family, medical care, and economic opportunity.
These are not edge cases. These are failures of basic dignity that have persisted across administrations, across congressional sessions, and across decades of advocacy. The people who have fought the longest for these things deserve honesty about where we are.
What Advocacy Actually Looks Like Right Now
If you are a non-disabled person reading this and wondering what you can do, the answer is not to share an infographic. The answer is to take a concrete action.
Fix something in your organization’s digital products
Not planning to fix it. Not adding it to the backlog. Fix something this week. Start with your most-visited pages and run them against the WCAG 2.1 Level AA Quick Reference. Bonus points if you add steps to make sure that kind of bug does not recur.
Review something that you have procured
Use the broader definition of procure, “buy, build, or use,” that I advocate for in my chapter on procurement in the recently published Digital Accessibility Ethics: Disability Inclusion in All Things Tech. If your organization buys software, your procurement contracts should include accessibility requirements and mechanisms for vendor accountability. Review your active contracts. If accessibility language is absent, add it at the next renewal.
Hire disabled accessibility professionals
The disability community is not a resource for inspiration. It is a talent pool with direct, lived experience of the barriers your products create, and that talent pool is unemployed at a rate more than double the general unemployment rate. Organizations that hire disabled professionals in accessibility roles produce better outcomes. It is not complicated.
Push back on people who want to continue delays
If your legal team is telling you to wait and see on Title II compliance, ask them how the litigation risk of inaction compares with the cost of remediation. The answer may surprise them. Remind them that Title II only sets the standard for measuring accessibility, and that the obligation to be accessible and provide effective communications already exists.
Demand accessible output from your AI coding tools.
Ask your development teams whether the AI-generated code they are shipping has been reviewed for accessibility. If it has not, establish a review process. The tools will not fix themselves.
We Cannot Declare Victory Until We Do Not Need GAAD
GAAD was founded in 2012. Fifteen GAADs later, the web remains largely inaccessible. Federal rules are under attack. AI is introducing new defects faster than we can remediate the old ones. The foundational protections of Section 504 required a nationwide mobilization just to hold in place. Medicaid and SNAP cuts threaten the independent living infrastructure that people with significant disabilities depend on to survive. The mental health system remains inaccessible, underfunded, and under-enforced. Some of the most fundamental accommodations people with disabilities have been promised remain years or decades away.
My biggest GAAD celebration will be the year it becomes unnecessary. I will celebrate it when screen reader users can navigate the web without hitting a wall of unlabeled buttons. When a wheelchair user in New York can take any subway to work. When a power wheelchair user can board a plane without transferring. When government digital services are accessible to everyone who needs them by default, because that is simply what government services are. When the safety net that allows people with disabilities to live in the community is treated as a basic civil right, not a budget line item subject to being slashed on a moment’s notice.
We are not there. Until we are, GAAD should not be viewed as a celebration. It is a painful reminder that people with disabilities still do not have the equal access we are supposedly entitled to.