Comparative Analysis of Generative Language Tools for Higher Education

The development of technology has marked great changes in all areas, especially in education, the incorporation of various tools supported by artificial intelligence has brought positive and negative consequences, in many cases there is a lack of knowledge of these and therefore they have not been properly used. This study attempts to evaluate and compare some generative language tools through a literature search and their practical use. Using content analysis, the seven selected tools were compared considering the criteria of: Efficiency, Quality, Usability, Scalability and Costs. The results show that tools such as ChatGPT, GitHub Copilot, Whisper, Hugging Face and Grammarly stand out for their high efficiency, quality, usability and scalability, being the most robust options for various educational applications. While tools such as Google Bard and Khan Academy's AI Tutor offer good evaluations in usability and scalability, with relatively low costs, making them accessible and effective in educational environments. Some limitations were identified in tools such as Talk to Books and ChatPDF that present challenges in efficiency and usability, limiting their applicability compared to more advanced tools. In addition, GPT for Slides, although functional for presentations, shows limitations in scalability and security. It was identified that there is a positive correlation between efficiency and quality in most tools, although there are exceptions where these aspects are not aligned. It is concluded that generative tools in Higher Education are crucial to modernize teaching and are aligned with the UN Sustainable Development Goals. Choosing the right tool should consider factors such as security, costs and ease of use to maximize its educational impact.