Acessibility in Generative AI Platforms: A Comparative Analysis of Popular Tools

• Arthur Diniz de Souza Zayed University http://orcid.org/0009-0007-6534-740X
• Erica de Castro Silveira Federal University of Ceará https://orcid.org/0009-0009-4471-5706
• Ingrid Teixeira Monteiro Federal University of Ceará https://orcid.org/0000-0001-5468-0724

DOI: https://doi.org/10.5753/ihc_estendido.2025.13268

Abstract

Introduction: Generative AI platforms, such as ChatGPT, Copilot, Gemini, and Claude, are widely adopted worldwide, yet little is known about their accessibility for people with disabilities. Objective: The objective of this study is to analyze the accessibility of web interfaces of leading generative AI platforms, identifying compliance and violations with the Web Content Accessibility Guidelines (WCAG) 2.1 and 2.2. Methodology: Automated accessibility evaluations were performed using the “Accessibility Insights for Web” tool, which internally applies the axe-core 4.10 engine to detect accessibility issues in the selected platforms. Results: All platforms evaluated presented at least one critical accessibility violation, which may hinder usage by people with disabilities. Based on these fndings, a set of best practices is proposed to guide developers in promoting digital inclusion and regulatory compliance.

Keywords: Accessibility, Generative AI, Web Interfaces, WCAG, Automated Validation

References

Acosta-Vargas, P. G., Guillén, M. L., e Quezada, A. M. (2024). Web accessibility evaluation of fifty ai applications: Empirical findings and challenges. International Journal of Human–Computer Interaction.

Brajnik, G. (2009). Validity and reliability of web accessibility guidelines. In Proceedings of the 2009 International Cross-Disciplinary Conference on Web Accessibility (W4A ’09). ACM.

Deque Systems Inc. (2025). axe-core: Accessibility Testing Engine, versão 4.10. Deque Systems. Disponível em: [link]. Acesso em: 09 jun. 2025.

Microsoft Accessibility Insights Team (2025). Accessibility Insights for Web, versão 2.46. Microsoft Corporation. Disponível em: [link]. io. Acesso em: 09 jun. 2025.

Sarangam, R. (2024). Screen-reader compatibility in generative chatbots: An exploratory study. In Proceedings of the 2024 ACM Web Conference. ACM.

Vigo, M., Brown, J., e Conway, V. (2013). Benchmarking web accessibility evaluation tools: Measuring the harm of sole reliance on automated tests. In Proceedings of the 10th International Cross-Disciplinary Conference on Web Accessibility (W4A ’13). ACM.

World Wide Web Consortium (2023). Web content accessibility guidelines (wcag) 2.2. Technical report, W3C. Disponível em: [link]. Acesso em: 09 jun. 2025.