Generative AI refers to a subset of artificial intelligence models that generate meaningful content in various forms, such as text, images, and multimedia. The Generative AI industry has witnessed rapid breakthroughs in recent years, with significant attention focused on large language models (LLMs). This paper aims to evaluate and assess the capabilities of Generative AI in generating supervised machine learning models, specifically focusing on four chosen LLMs: OpenAI’s ChatGPT, Anthropic’s Claude, Google’s Gemini, and Microsoft’s Copilot. This study employs a multi-phase methodology, where models are trained on various algorithms using traditional methods. Then, the LLMs followed a multi-step prompting strategy, emulating the traditional machine learning pipeline. The LLMs were evaluated based on their responses to direct prompts and their ability to generate Python code. The results show that Generative AI is capable of generating supervised machine learning models, with certain LLMs performing better than others. Claude excelled in preprocessing data, while Gemini performed best in training and testing machine learning models. Overall, there is still room for improvement. The models occasionally panicked and failed to train in a considerable number of instances. These findings could pave the way for further research on the potential contributions of Generative AI to the field of machine learning. © 2026, Springer Science and Business Media Deutschland GmbH. All rights reserved.

https://www.scopus.com/pages/publications/105040601598?origin=resultslist