Presentation of the special issue “Generative AI, ChatGPT and Education. Consequences for Intelligent Learning and Educational Evaluation”

Abstract

In July 2023, given the rise of LLMs (Large Langauge Models), RED convened this special issue on Generative AI and Education, where special attention was paid to its consequences for intelligent learning and educational evaluation.

We wanted to give space to contributions that included research related to these topics. Also to experiences about intelligent learning and formative evaluation in ChatGPT contexts.

The call was with these general questions

• Does AI have the potential to revolutionize existing teaching methods, assessment and student support?
• Creative thinking and problem solving are essential in modern and very complex environments. Could this AI help students deal with these problems?

We also had doubts about its benefits. They could be summarized in this question: Generative AI will begin to serve as an active partner in social, creative and intellectual actions continuously over time, and not only as an answer to isolated questions: What are the impacts that will occur? Now those impacts are unknown in the practices that may exist.

Another intention was:

A theoretical framework is needed to address these questions and in general for an effective deployment of AI systems in education. It is necessary to do so beyond the results provided by empirical research. And that it does not guide and direct at new crossroads, both in research and practice.

In the conclusions we see to what extent these expectations have been met. As a consequence, we deduce that the critical importance of theory in the design, development and deployment of AI in education is necessary now more than ever. But we are equally underserved.

In this perspective, we continue to critically consider the relevance and continuity of existing learning theories when AI becomes a reality in classrooms.

As that result is not met, we also reiterate the call to consider new frameworks, models and ways of thinking. We are referring to those that include the presence of non-human agents, which we hesitate to call a new technology, because it is more like an active partner than a simple technology, as has happened until now.

This approach is precisely what makes us insist on a series of important questions for the future, precisely about the review of learning theories based on existing configurations. And to investigate what their alternatives would be in this case.

We have done after extensive and exhaustive dissemination in your call. But, despite this and beyond these general conclusions that we have made, the special issue offers us evidence of a scarce empirical investigation of practical cases in the application of generative AI in education.

However, of the hundred or so contributions received, seven have been selected in the previous editorial review. The rest have been discarded because they do not conform to the standards or are not the type of contributions requested (the literature reviews per se and the self-report studies stand out among them, due to their high number).

Of those seven, six have passed editorial review. They are described at the end.

The main contributions of this small number of contributions have been the confirmation of a low level of research and practice. Also, some very interesting contributions from the articles and essays by the invited authors.

We draw your attention to these articles and the clear results and evidence obtained on the concrete use of generative AI in specific environments. Results of inevitable use by schools, universities and teachers in these environments or in others to which they can be transferred.