AI AS AN INTERACTIVE AGENT IN THE TEACHING OF THE HISTORY OF THE MILITARY DICTATORSHIP IN BRAZIL

Techniques and Possibilities

Abstract

The generative Artificial Intelligence devices for text have been increasingly used by students and teachers since the popularization of ChatGPT in 2022. Issues such as ethics and authorship have been pointed out as “machine weaknesses”. However, they pertain to individual and social behaviors in the use of technological devices, revealing “human weaknesses” in digital handling. In the field of historical science, the teacher’s craft requires a critical perspective not only on the content but also on the techniques for accessing data available online, involving processes of identification, selection, critical analysis, and decision-making regarding information. Just like textbooks, films, and newspapers, chatbots emerge as tools for developing creativity, autonomy, and interactivity. Their critical evaluation is necessary, as are the human decision-making principles that govern their use. Therefore, this article presents a contextualized practice of using ChatGPT to create a storytelling about the period of the civil-military dictatorship, starting in 1968, featuring a historian and a former guerrilla fighter as protagonists, as a starting point for teaching and learning about the topic. The article aims to extend the experiment conducted by historian Julio Bentivoglio (2023), building on a distinct dialogical configuration in which fiction and contextual reality are intertwined to enable the comprehension of history. To this end, it draws on research on AI, Machine Learning, and education through digital technologies, as well as studies on the civil-military dictatorship.