Banca de DEFESA: DANIEL OLIVEIRA DOS SANTOS

Uma banca de DEFESA de MESTRADO foi cadastrada pelo programa.
STUDENT : DANIEL OLIVEIRA DOS SANTOS
DATE: 27/03/2025
TIME: 14:00
LOCAL: SALA DE ATOS DA FE 1 - HÍBRIDO - RNP
TITLE:

PERCEPTIONS OF EDUCATION AND HEALTH TEACHERS ON THE RISKS, BENEFITS AND FEASIBILITY OF ADOPTING HYBRID TEXTS WRITTEN BY STUDENTS AND ARTIFICIAL INTELLIGENCE IN UNDERGRADUATE RESEARCH

KEY WORDS:

Artificial Intelligence; academic writing; hybrid texts; higher education

PAGES: 137
BIG AREA: Ciências Humanas
AREA: Educação
SUBÁREA: Fundamentos da Educação
SPECIALTY: Sociologia da Educação
SUMMARY:

The research investigates the understanding of faculty members in the fields of education and health about the use of Artificial Intelligence (AI) technologies in the production of hybrid academic texts by their students. In contextualizing the research, aspects of the hybrid text and academic integrity in higher education are discussed. The methodology employed follows a qualitative approach, with content analysis based on Bardin (1977) and data triangulation according to Triviños (1987). Data collection was conducted through ten semi-structured interviews with faculty members at a private university in the Federal District. Nine out of the ten interviewed professors have already used generative AI tools, albeit in an exploratory manner. Among the various perceptions about AI in textual production, the faculty's understanding of the risks, benefits, and conditions to enable the adoption of hybrid text practices (AI and students) in undergraduate research stands out. The research results indicated the "risks" of AI in text writing are associated with the following subcategories and mentions in the interview content: plagiarism [05], ethics and morality [05], lack of critical reading [04], dependence on AI [02], and immediacy [02]. The notion of risk is linked to negative perceptions about the use of AI in writing, with a dominant alignment regarding the "negative" understanding of the students' ability to produce hybrid texts, represented by the subcategories plagiarism [05], ethics and morality [04], defining differentiation in the hybrid text: AI and human [04]. The "benefits" were associated with the subcategories assistance in writing [06], idea generator/inspirer [03], augmented intelligence [02]; confronting concepts (of AI and human) [02]; assistance in text interpretation [01]; personalization [01]. The conditions to enable the adoption of hybrid text practices (AI and students) are associated with training [04], ethics and morality [02], regulation of AI use [02]; guidance [01]. Aspects that highlight conditionalities for the use of AI in undergraduate and those that are ambivalent, are represented, respectively, by the need for training and the paradoxical sense between qualifying mature students versus immature ones for the appropriation of AI in writing hybrid texts [03]. To mitigate the risks of misuse of AI and its negative effects, the "conditional" category with training [05] and "normative": regulation [3] were evidenced. The research highlights that professors face difficulties in understanding what is beneficial in the use of AI. The faculty still do not have a systematic and routine practice of using generative AI, being in a phase of pre-appropriation of this technology, which requires qualification processes. Additionally, there is an urgent need for standardization that establishes guidelines and ways of using AI in the university context.

COMMITTEE MEMBERS:
Interno - 1651994 - CARLOS ALBERTO LOPES DE SOUSA
Externa ao Programa - 1957211 - CANDIDA BEATRIZ ALVES - UnBExterno ao Programa - 2662431 - EDUARDO DI DEUS - UnBExterna à Instituição - GRACIELLA WATANABE