Banca de DEFESA: TEREZA CRISTINA NUNES DE PAULA DA SILVA

Uma banca de DEFESA de MESTRADO foi cadastrada pelo programa.
STUDENT : TEREZA CRISTINA NUNES DE PAULA DA SILVA
DATE: 18/07/2025
TIME: 09:30
LOCAL: SALA DE ATOS DA FE 1 - HÍBRIDO - RNP
TITLE:

Will Everyone Have Their Own Pet Artificial Intelligence (AI)?:

Teachers’ Perceptions of the Anthropomorphization of AI in the Hybrid Text Production by 5th Grade Elementary Students

KEY WORDS:

Artificial Intelligence; Teacher Perception; Hybrid Text; Anthropomorphization; Education; Early Years of Elementary School

PAGES: 113
BIG AREA: Ciências Humanas
AREA: Educação
SUBÁREA: Fundamentos da Educação
SPECIALTY: Sociologia da Educação
SUMMARY:

The objective of this research is to critically analyze, from a prospective standpoint, teachers' perceptions regarding the appropriation of generative AI in writing tasks during the early years of elementary education—particularly in the writing of texts in relation to the trend of anthropomorphizing technology by 5th-grade students. Based on the concept of hybrid written text proposed by Lopes, Forgas, and Cerdá-Navarro (2024), the study integrates research on the use of AI in textual production (Dugan et al., 2023; Lopes, Forgas, & Cerdá-Navarro, 2024; Zeng et al., 2024), as well as studies on the anthropomorphization trend of generative AI (Placani, 2024; Barrow, 2024; Kalemkuş & Kalemkuş, 2025). This study is grounded in the conception of education as a liberating, dialogical, and critical process, as defined by Freire (2023). Employing a qualitative, descriptive, and exploratory approach, the methodological procedures are based on Bardin’s (1977) content analysis and Triviños’ (1987) data triangulation method. Data collection was conducted through research instruments that included a questionnaire answered by 56 respondents and a semi-structured interview carried out with 16 participants. The findings highlight the thematic categories: “contributions of AI to textual production,” “harmful effects of AI on writing,” and “human dimension in the use of AI in textual production,” all in the context of generative AI appropriation in the early years of schooling. Data from the questionnaire show that 66.1% of respondents either strongly agreed or agreed that the use of AI in writing tasks leads to student dependency on technology. Coupled with the categories “teachers' appreciation of students' capacity to produce hybrid texts” and “conditions required for hybrid text production,” it is inferred that teachers do not accept the appropriation of AI by 5th-grade students at the student’s sole discretion unless it is embedded within a pedagogical process and guided by teacher mediation. The questionnaire data reveal that the combined agreement level—between “strongly agree” and “agree”—regarding the idea that the use of AI by 5th-grade students hinders writing development reached 67.9%. It is evident that what teachers describe as the “paper experience”—a moment of writing performed by the student without technological assistance—should be prioritized in the early years of schooling. The category “human dimension in the use of AI in textual production” leads to the conclusion that, even when AI is used, authorial aspects must be present in the writing process, acknowledging the singularity, historical context, 

and historicity of each subject involved in text production. The category “risks of AI anthropomorphization” contributes to examining teachers’ perceptions regarding the personification of generative AI by 5th-grade students in the production of hybrid texts and other technological uses. Prospectively, it is noted that when students produce a hybrid text, they tend to anthropomorphize and attribute human characteristics to generative AI. The trust bond with the machine; the notion of AI as a pet or companion; AI as pleasant; technological interaction prevailing over human interaction; AI as a facilitator; and AI as seductive—all represent the risks of anthropomorphizing generative AI in the context of textual production and other technological uses by 5th-grade students. It is inferred that, if not addressed responsibly, the implications of anthropomorphization hinder a comprehensive understanding of issues such as data usage, trust in systems, information homogenization, and other factors involved in generative AI usage. The interaction between students and generative AI systems opens new avenues for future research, as attributing emotions to machines may result in emotional and social challenges for human beings and historicity of each subject involved in text production. The category “risks of AI anthropomorphization” contributes to examining teachers’ perceptions regarding the personification of generative AI by 5th-grade students in the production of hybrid texts and other technological uses. Prospectively, it is noted that when students produce a hybrid text, they tend to anthropomorphize and attribute human characteristics to generative AI. The trust bond with the machine; the notion of AI as a pet or companion; AI as pleasant; technological interaction prevailing over human interaction; AI as a facilitator; and AI as seductive—all represent the risks of anthropomorphizing generative AI in the context of textual production and other technological uses by 5th-grade students. It is inferred that, if not addressed responsibly, the implications of anthropomorphization hinder a comprehensive understanding of issues such as data usage, trust in systems, information homogenization, and other factors involved in generative AI usage. The interaction between students and generative AI systems opens new avenues for future research, as attributing emotions to machines may result in emotional and social challenges for human beings.

COMMITTEE MEMBERS:
Interna - 1151554 - ANDREA CRISTINA VERSUTI
Interno - 1651994 - CARLOS ALBERTO LOPES DE SOUSA
Externa ao Programa - 1957211 - CANDIDA BEATRIZ ALVES - UnBExterna à Instituição - MARCIA LOPES REIS - UNESP