Mariusz Jarocki, Małgorzata Kowalska-Chrzanowska
The development of artificial intelligence (AI) between 2022 and 2025 has led to the emergence of numerous tools on the IT market designed to support the creation of websites. Some of these are geared towards closed platforms (WiX, Squarespace), whilst others are dedicated to open-source content management systems (CMS). This paper focuses on an analysis of themes created using AI tools for the three dominant open-source CMS platforms on the market: WordPress, Joomla and Drupal.
The study analysed both specialist solutions, available only for specific platforms, and universal AI models. When comparing the generated themes, the focus was on assessing the quality of the generated layout, responsiveness, compliance with the standards of the given platform, compliance with WCAG guidelines, and the correctness of integration with the CMS.
The results indicate a significant dominance of solutions designed for the WordPress system, not only in terms of the number of dedicated projects but also their greater maturity; support for the Joomla and Drupal platforms proved to be patchy. It was also demonstrated that large generative language models (ChatGPT, Gemini, Claude) are capable of correctly generating code snippets for the CMSs under investigation, which can successfully serve as a basis for building complete themes.