Leveraging AI and Emerging Technologies for Accessibility
Wins, Warnings, and What's Next
Terisa O'Dowd, PhD�Instructional Accessibility Designer�Texas Woman's University
Your Experience with AI
Interactive Poll
Raise your hand if you’ve used an AI tool recently. �Keep it raised if you’ve used AI specifically for accessibility?
Mentimeter poll:
Why This Matters Now
AI is Transforming Education
Artificial intelligence has become deeply integrated into educational platforms, learning management systems, and accessibility tools.
Understanding its capabilities and limitations is essential for creating truly inclusive learning environments.
85%
Growth in AI tools
Used in education since 2023
Our Journey Today
01
Wins
Where AI excels in �accessibility
02
Warnings
Where AI falls short and risks emerge
03
What's Next
Human + AI collaboration for the future
AI Wins for Accessibility
Alt Text Generation
Automated descriptions for images and graphics
Auto-Captioning
Real-time transcription and subtitle creation
Text Simplification
Complex content made clear and digestible
Testing Automation
Rapid accessibility compliance scanning
Demo: AI-Generated Alt Text
Tools to Explore
What to Compare
AI-generated descriptions vs. human-written alt text
Accuracy, context, and relevance in educational materials
Live demonstration: Watch how AI interprets visual content and generates descriptive text
Test your “image description” skills against AI
How would you describe this image?
Demo: AI-Powered Captioning
Otter.ai�
Microsoft Word Translation
Microsoft Stream
Watch real-time captioning in action and explore accuracy across different speaking styles and environments
Demo: Text Simplification & Summarization
Complex Academic Text
Dense terminology and lengthy�explanations
AI Processing
ChatGPT, TLDR This, or �NotebookLM
Accessible Content
Clear, concise, and �comprehensible
See how AI can simplify complex course materials while preserving essential meaning for diverse learners
Complex Text:
The inexorable march of technological progress, specifically the pervasive integration of artificial intelligence within pedagogical paradigms, fundamentally necessitates a recalibration of entrenched educational methodologies; the inherent didactic efficacy of conventional instructional modalities is arguably diminished in an epoch characterized by algorithmic ubiquity and instantaneous data accessibility.
Demo: AI-Assisted Accessibility Testing
Automated Scanning Tools
Watch how AI rapidly identifies accessibility issues across documents and web content
Keyboard test
Go to any AI tool you have access to and see if you �can interact and engage with the tool using only your keyboard.
For example, ChatGPT: https://chatgpt.com
Interactive Brainstorm
Share Your Challenge
What accessibility barrier do you face most often in your work?
Dream Solution
What do you wish AI could do to solve this accessibility barrier?
Join the live Slido poll to contribute your ideas and see what others are thinking
Menitmeter word cloud question
AI Risks and Red Flags
Context Errors
AI misses nuance, cultural references, and discipline-specific meanings
Algorithmic Bias
Training data can perpetuate stereotypes and exclude marginalized voices
Privacy Concerns
Student data and educational materials shared with third-party AI services
Inaccessible AI Tools
Many AI platforms lack proper keyboard navigation and screen reader support
Demo: Chatbot Accessibility Check
1
Keyboard Navigation Test
Can users access all features without a mouse?
2
Screen Reader Compatibility
Are responses properly announced and navigable?
3
Visual Contrast Check
Does the interface meet WCAG color requirements?
Test popular AI chatbots like ChatGPT and Claude for accessibility compliance
Ethical Scenarios: Your Decision
1
Auto-Captions Only
A faculty member uses auto-generated captions without reviewing them. Is this sufficient for accessibility compliance?
2
Free AI Tools with Student Data
An instructor uploads student work to a free AI service for feedback. What are the privacy implications?
3
Missing Alt Text
AI generates generic descriptions like "an image" for complex diagrams. Should this be published as-is?
Discussion prompt: What would you do in each scenario? What policies should guide these decisions?
The Future: Human + AI Collaboration
Accuracy
Always verify AI output
Human Oversight
Maintain expert review
Accessibility
Ensure tools are inclusive
Bias Awareness
Check for representation
Privacy
Protect student data
Five principles for responsible AI use in accessibility work.
Take Action: Next Steps
What's Next in AI & Accessibility
Credit: ProStockStudio / Shutterstock.com © 2024
Let's Stay Connected
Thank you for joining today!
Continue the conversation
📧 Email: todowd1@twu.edu
💼 LinkedIn: terisaodowd
Final Poll: How likely are you to experiment with AI for accessibility after today's session? Share your thoughts on Terisa’s LinkedIn.