A company redesigned its website with AI features.
AI chatbot for customer service. Dynamic content that personalizes based on user behavior. Smart navigation that adapts to preferences.
The site looks modern. Performance metrics are excellent. Customer satisfaction scores went up.
Then the demand letters arrived from disability rights lawyers.
The AI features don’t work with screen readers. Blind users can’t access the chatbot. Dynamic content creates navigation barriers. The personalization breaks assistive technology.
The company violated the Americans with Disabilities Act. Now they face a class action lawsuit and DOJ investigation.
And the development team had no idea AI features could create accessibility violations.
How AI Breaks Accessibility
Companies think AI makes websites better for everyone.
For users with disabilities, AI often makes websites unusable.
AI chatbots without accessibility: Most AI chatbots aren’t coded for screen reader compatibility. Blind users can’t navigate the chat interface, can’t tell when AI responds, and can’t access chat history or controls.
Dynamic AI content: Content that changes based on AI recommendations confuses screen readers. Users can’t predict page structure or find information consistently.
AI-powered navigation: Personalized menus and adaptive interfaces break the predictable navigation that assistive technology users depend on.
AI image descriptions: Automated alt text from AI often misses context or describes images incorrectly, making visual content inaccessible.
Each AI feature that improves experience for some users creates barriers for others.
The ADA Applies to Websites
The Americans with Disabilities Act requires equal access to places of public accommodation.
Courts have consistently found that websites are places of public accommodation subject to ADA.
What this means: Companies must make websites accessible to people with disabilities including those who are blind, deaf, have mobility impairments, or cognitive disabilities.
How it’s measured: Courts use Web Content Accessibility Guidelines (WCAG) as the standard for ADA compliance. WCAG 2.1 Level AA is generally required.
AI features must be accessible: When companies add AI to websites, those features must meet the same accessibility standards as everything else.
Most companies deploy AI website features without any accessibility testing.
Real Violations Companies Are Making
The inaccessible chatbot: Company adds AI chatbot to homepage. The chat widget isn’t keyboard accessible. Screen readers can’t announce new messages. Blind users can’t use customer service.
The dynamic content problem: AI personalizes page content based on user behavior. Content shifts location unpredictably. Screen reader users can’t find consistent navigation landmarks.
The broken forms: AI assists with form completion and validation. Error messages generated by AI aren’t properly labeled. Users with screen readers can’t tell what went wrong or how to fix it.
The video caption failure: AI generates captions for video content. The captions are inaccurate and miss important context. Deaf users get wrong information.
The image description disaster: AI creates alt text for images. The descriptions are vague, incorrect, or unhelpful for understanding visual content.
Each of these violates WCAG standards and creates ADA liability.
The Lawsuit Wave
ADA website accessibility lawsuits are increasing, and AI features are becoming a focus.
Demand letters: Disability rights lawyers send demand letters claiming ADA violations. They threaten litigation unless the company pays settlement and fixes accessibility.
Class actions: Lawsuits on behalf of all users with disabilities who can’t access the website. Potential damages multiply by number of affected users.
DOJ investigations: Department of Justice investigates systemic accessibility violations and can impose consent decrees requiring extensive remediation.
Injunctive relief: Courts order companies to make websites accessible, which can be expensive and time-consuming with AI features.
Companies that never considered accessibility are discovering their AI features created massive liability.
Why “We Didn’t Know” Doesn’t Help
Companies argue they didn’t realize AI features would break accessibility.
This defense rarely works.
Constructive knowledge: Companies should know ADA applies to websites. They should know AI features must be accessible.
Industry standards: WCAG guidelines exist. Companies are expected to follow them.
Willful blindness: Deploying AI without accessibility testing shows deliberate indifference to disability rights.
Prior notice: If the company received any accessibility complaints before, they had clear notice of the requirement.
Ignorance of accessibility requirements doesn’t eliminate liability.
What WCAG Compliance Requires
Web Content Accessibility Guidelines set specific technical standards.
For AI chatbots: Keyboard navigation without mouse required, screen reader announcements for new messages, clear labels for all controls, and alternative methods to access the same information.
For dynamic content: Consistent navigation structure, notification when content changes, ability to pause or control automatic updates, and predictable behavior.
For AI-generated content: Accurate alt text for images, proper heading structure, meaningful link text, and form labels that screen readers can interpret.
For multimedia: Accurate captions for audio, audio descriptions for visual content, transcript availability, and no reliance on color alone to convey meaning.
Most AI features violate multiple WCAG requirements.
What Companies Should Do
If companies have deployed or are planning AI website features, accessibility must be addressed.
Test with assistive technology: Use screen readers, keyboard-only navigation, and other assistive tools to test AI features before launch.
Follow WCAG guidelines: Ensure AI features meet WCAG 2.1 Level AA standards at minimum.
Involve users with disabilities: Get feedback from people who actually use assistive technology.
Provide alternatives: When AI features can’t be made fully accessible, offer equivalent alternatives.
Monitor and fix: Continuously test accessibility as AI features evolve and fix issues promptly.
Train development teams: Educate developers on accessibility requirements for AI features.
Most companies do none of this because they assume modern AI features are automatically accessible.
The Remediation Cost
Fixing accessibility after launch is expensive.
Redesign costs: Rebuilding AI features for accessibility takes significant development time.
Testing costs: Proper accessibility testing requires specialized expertise and tools.
Legal costs: Defending lawsuits and negotiating settlements adds up quickly.
Ongoing monitoring: Maintaining accessibility as AI evolves requires continuous effort.
Building accessibility in from the start costs far less than fixing it under legal pressure.
Ask Yourself This
Does your website have AI chatbots, dynamic content, personalized experiences, or other AI features?
Have those features been tested with screen readers and other assistive technology?
Can users with disabilities fully access and use the AI features?
If you don’t know the answers, there’s an ADA compliance problem waiting for a demand letter to expose it.
Schedule a 30-minute AI Accessibility Risk Assessment where we’ll:
- Review what AI website features may violate the ADA
- Show what WCAG compliance requires for AI
- Identify current accessibility gaps
- Create a remediation plan for AI features
- Help implement accessibility testing processes
Adding AI to your website feels like innovation until the ADA lawsuit reveals blind users can’t access any of it.
Brian Alexander
Founding Partner, VisionAI+ Consulting Group
Former General Counsel, AI Engineering Firm
Certified ISO 42001 Lead Auditor