The premise
AI can draft an access-focused review checklist that catches AI workflows that exclude disabled employees or customers.
What AI does well here
- Check whether output is screen-reader-friendly
- Test plain-language and reading-level alternatives
- Surface assumptions about modality (voice, vision, motor)
What AI cannot do
- Replace user testing with disabled people
- Audit the underlying model
- Make the accessibility decision
End-of-lesson check
15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-ethics-ai-disability-access-prompt-review-creators
What is the PRIMARY purpose of using AI to review internal AI prompts for disability access?
- To test the underlying AI model's technical accuracy
- To generate compliance documentation for regulators
- To replace manual accessibility reviews entirely
- To catch workflows that exclude disabled employees or customers
Which of the following is an example of an accessibility issue AI can help identify in AI-generated outputs?
- Whether the output is grammatically correct
- Whether output is screen-reader-friendly
- Whether the color scheme matches brand guidelines
- Whether the writing tone matches company voice
Which assumption about user modality would MOST likely create an accessibility barrier for users with motor impairments?
- Assuming users can use a mouse or trackpad
- Assuming users can see graphical elements
- Assuming users can hear audio
- Assuming users can read plain text
An AI-powered customer service chatbot only responds with voice audio. Which assistive technology would be MOST directly excluded?
- Wheelchair accessibility features
- Hearing aids for deaf users
- Braille displays
- Screen readers for blind users
Why should AI-generated test cases involving assistive technology NOT replace user testing with disabled people?
- AI-generated test cases are too expensive to create
- Regulations prohibit AI from simulating disability scenarios
- AI cannot replicate the lived experience of using assistive technology daily
- Disabled users cannot provide reliable feedback
What does the WCAG acronym stand for in accessibility contexts?
- Website Compliance and Accessibility Gateway
- World Center for Accessible Design
- Wide Computer Access Governance
- Web Content Accessibility Guidelines
When reviewing an AI prompt for accessibility, which issue relates to assumptions about vision?
- Generating complex vocabulary without plain-language alternatives
- Requiring fine motor control for interactions
- Displaying results only as images without text descriptions
- Requiring voice input without keyboard alternatives
What is one capability AI CANNOT provide when reviewing prompts for disability access?
- Generating test cases for common assistive tech
- Auditing the underlying AI model for bias
- Testing reading level alternatives
- Checking if outputs are screen-reader-compatible
A workflow generates complex financial reports. Which AI accessibility check would MOST improve access for users with cognitive disabilities?
- Ensuring keyboard navigation works
- Adding voice output for all text
- Generating plain-language summaries alongside technical content
- Adding color contrast to charts
What type of accessibility issue involves 'reading level' when reviewing AI outputs?
- Whether charts include data tables as alternatives
- Whether the interface supports text-to-speech
- Whether content uses simple vocabulary and sentence structure
- Whether text appears in a readable font size
Which statement best describes the role of AI in making final accessibility decisions?
- AI decisions override all other accessibility considerations
- AI should make the final accessibility decision based on its analysis
- AI can suggest issues, but humans must make accessibility decisions
- AI decisions should be verified by another AI system
An internal AI system generates image descriptions for a document management system. What accessibility feature should be verified?
- Whether images match brand colors
- Whether images are high resolution
- Whether images have alternative text
- Whether images are compressed for faster loading
What does AI do WELL when used for disability access review of prompts?
- Guarantees WCAG compliance automatically
- Surfaces assumptions about modality that might exclude users
- Determines if an AI system is unbiased
- Replaces all accessibility testing with disabled users
A voice-activated AI assistant assumes users can speak clearly into a microphone. This primarily excludes users with which disability?
- Visual impairments
- Hearing impairments
- Speech or motor impairments that affect voice
- Cognitive disabilities
When creating test cases for assistive technology compatibility, which approach follows best practices?
- Testing only visual accessibility
- Testing only with the most common screen reader
- Testing with multiple assistive technologies and real users
- Relying solely on automated accessibility checkers