Overcoming challenges of web accessibility: Automated testing in CI/CD pipelines and AI-powered remediation

Thumbnail Image

URL

Journal Title

Journal ISSN

Volume Title

School of Science | Master's thesis

Authors

Date

2025-10-05

Department

Major/Subject

Computer Science

Mcode

Degree programme

Master's Programme in Computer, Communication and Information Sciences
Master's Programme in Computer, Communication and Information Sciences
Master's Programme in Computer, Communication and Information Sciences

Language

en

Pages

60

Series

Abstract

Web accessibility refers to the quality of a web service that allows access and usability for people irrespective of their abilities, emphasizing the needs of disabled individuals. Recently, automatically ensuring accessibility for web services has become an important requirement that needs a practical approach to achieve. Continuous Integration and Continuous Delivery (CI/CD) can help to automate and encapsulate accessibility tests in the automated quality assessment process during the software life cycle. This method can dramatically diminish the time and effort web developers spend testing accessibility in comparison with other manual approaches. In addition, using the CI/CD pipeline also helps to consistently maintain the accessibility level of a web service throughout the development process. However, practically integrating accessibility testing into the pipeline encounters several problems, especially with services not being evaluated for accessibility. Moreover, Large Language Models (LLMs) can be used to effectively support developers in resolving the accessibility violations that are discovered by the automated tools. This thesis identifies the advantages of adopting accessibility tests in CI/CD pipelines and analyzes the challenges of applying this approach in a practical context. The research reveals that a large number of accessibility issues cause major difficulties in the integration of accessibility assessment into the CI/CD pipeline. Those issues need to be continuously addressed without affecting the workflow of the current CI/CD pipeline. This thesis also proposes and evaluates a practical solution to address the existing challenge by implementing a framework to effectively maintain web accessibility using the CI/CD pipeline and AI Agent.

Description

Supervisor

Vuorimaa, Petri

Keywords

web accessibility, accessibility testing, automated testing, CI/CD, large language model, AI agents

Other note

Citation

Permanent link to this item

https://urn.fi/URN:NBN:fi:aalto-202512179373

Collections

[dipl] Perustieteiden korkeakoulu / SCI