LLM-assisted coding tools have accelerated the pace of software development. Unfortunately, accessibility compliance has taken a back seat in the process. Here’s how to address the problem.

The New Stack is a publication covering trends and technologies in cloud-native development, DevOps, and software delivery. Developers can learn about containerization, Kubernetes, and cloud computing, as well as explore topics such as microservices architecture, serverless computing, and continuous integration/continuous delivery (CI/CD) pipelines.

The New Stack

LLMs trained on an inaccessible web are perpetuating accessibility failures in AI-generated code. Data from AudioEye and the 2026 WebAIM Million report shows 95.9% of top homepages have detectable WCAG failures, with the average page containing 297 accessibility issues. Common AI-generated problems include conflicting ARIA labels, incorrect heading hierarchy, and keyboard traps. Practical recommendations include integrating accessibility testing into the SDLC, using automated monitoring tools, avoiding overlay widgets, and educating developers on semantic HTML and WCAG compliance.

LLMs were trained on an inaccessible web — AudioEye data shows AI is still building one

Workflow: Where accessibility testing should live in your development process