DIGITAL LIBRARY
GENERATION OF LINEAR SYSTEMS EXERCISES USING ARTIFICIAL INTELLIGENCE
Universidad Autónoma Metropolitana (MEXICO)
About this paper:
Appears in: INTED2026 Proceedings
Publication year: 2026
Article: 1670
ISBN: 978-84-09-82385-7
ISSN: 2340-1079
doi: 10.21125/inted.2026.1670
Conference name: 20th International Technology, Education and Development Conference
Dates: 2-4 March, 2026
Location: Valencia, Spain
Abstract:
Whenever a course is taught in a higher education institution—such as Numerical Methods—exercises are required for classroom examples, assignments, or exam questions. A vast number of exercises is needed to prevent students from having the answers in advance and to discourage cheating. In turn, these exercises must include their respective solutions, which should be generated within a timeframe similar to that of an examination period.

The present work demonstrates the generation of exercises for the topic of systems of linear equations. AI has revolutionized practically every field of human knowledge, particularly the field of education. AI can generate different types of content: text, images, software, music, etc. Specifically, AI chatbots allow for the generation of texts on various topics and can be used to create exercises along with their solutions.

Several AI chatbots were tested: ChatGPT, Copilot, Gemini, and Perplexity. The plans for using these AI chatbots can be classified into two categories: Free and Pro. Both differ in terms of features and limitations. The exercises generated focus on systems of linear equations with an interpretation relevant to an academic context.

The results obtained from the same prompt across all considered chatbots are compared. Furthermore, results from Free versus Pro plans are evaluated within the topic of systems of linear equations for the Numerical Methods in Engineering course at UAM Azcapotzalco.
Keywords:
e-learning, artificial intelligence, problem-based learning, numerical methods, linear systems.