Improving Accessibility through Leveraging Large Language Models (LLMs)
summary
Large language models produce image descriptions, real‑time captions, translations, plain‑language rewrites and code suggestions, but require human review because hallucinations and IP issues occur; a local LLM is plugged into Jira to summarize tickets, map them to specific WCAG success criteria and generate VPATs, and an internal chatbot auto‑creates consistent error messages. Typical outcome: manual test plan takes two to four hours vs 20–30 minutes after LLM draft, so it speeds prep but not final validation.
AI-written summary · about
chapters (5)
- 0:00Introduction and background
- 10:10LLMs in accessibility applications
- 29:22Project implementation and use cases
- 32:47Limitations and future of AI in accessibility
- 35:33Q&A session
related talks
33mExplore LLMs and Multimodal Approaches in Test Automation
49mA Practical Approach to Using LLMs: Peculiarities and Pitfalls
43mAI Development Lifecycle: Learnings of What Changed with LLMs
29mAutomating Privacy Reviews with LLMs: Challenges & Opportunities
33mOpening Keynote: LLMs, between fantasy and reality - the poin...
47mValidating LLMs – A True Story
49mNext-Gen Documentation with AI
39mLLMs and Hallucinations: Understand, Measure, Control