Generative Artificial Intelligence for Controlled Content Generation

The main objective of this article is to provide an overview of how Generative Artificial Intelligence techniques can be used to create content in a controlled manner for applications that require high reliability and accuracy. Generative Artificial Intelligence, capable of generating various types of content, such as text, images, video, and audio, has revolutionised many areas of knowledge, but it must be considered that these systems can suffer from ’hallucinations’ and produce false or inconsistent information. For this reason, having a system that allows their controlled use in sensitive contexts is of vital importance. This paper presents a methodology that allows the user to use a Generative Artificial Intelligence system to create new content according to the user’s needs based on controlled and user-specified sources. The operation of the proposed system is illustrated by means of a concrete example of application in education, the automatic generation of test-type exams.

Authors

• J. M. Torres
• C. A. Martin
• M. A. Bacallado
• R. Aguilar
• S. Alayón

Keywords

• Generative Artificial Intelligence
• Retrieval Augmented Generation
• User-centred automation
• Education
• Exam generation