Abstract
Background
Recent advances in generative AI (genAI) have exposed students and teachers worldwide to tools with increasingly human-like capabilities. Such genAI tools lend themselves to anthropomorphism, the tendency to attribute humanlike characteristics, motivations, intentions, or emotions to non-human entities. Anthropomorphism can affect users’ experiences and learning. However, the role of anthropomorphism of genAI in higher educational settings remains unclear.
Objective
This poster presents a scoping review of empirical studies on LLM-based genAI tools in higher education since 2022 that address anthropomorphism. We investigate:
RQ1: What are the characteristics and methodological approaches of empirical higher education research examining genAI anthropomorphism since 2022?
RQ2: How is genAI anthropomorphism conceptualised in this research, and what are the key findings and implications regarding its role in higher education contexts?
Methods
Following the PRISMA-SCR guidelines, we searched 6 databases, screened 332 papers, and found 9 papers eligible for review.
Results
RQ1: The few existing studies were recent and geographically diverse, primarily student-focused, mostly ChatGPT-focused, and primarily used cross-sectional surveys, but with different measurement scales.
RQ2: Most studies examined anthropomorphic response, while only one study examined anthropomorphic cues. Four studies linked higher anthropomorphism to positive effects on trust and learning, one found a negative link to trust, and three found no significant impact on student preferences and learning.
The review found mixed evidence but suggests that anthropomorphism can significantly influence trust, engagement, and learning. We highlight the need for standardised anthropomorphism assessments and (quasi-)experimental designs to establish causality, observe long-term effects, and rigorously assess interventions.
Recent advances in generative AI (genAI) have exposed students and teachers worldwide to tools with increasingly human-like capabilities. Such genAI tools lend themselves to anthropomorphism, the tendency to attribute humanlike characteristics, motivations, intentions, or emotions to non-human entities. Anthropomorphism can affect users’ experiences and learning. However, the role of anthropomorphism of genAI in higher educational settings remains unclear.
Objective
This poster presents a scoping review of empirical studies on LLM-based genAI tools in higher education since 2022 that address anthropomorphism. We investigate:
RQ1: What are the characteristics and methodological approaches of empirical higher education research examining genAI anthropomorphism since 2022?
RQ2: How is genAI anthropomorphism conceptualised in this research, and what are the key findings and implications regarding its role in higher education contexts?
Methods
Following the PRISMA-SCR guidelines, we searched 6 databases, screened 332 papers, and found 9 papers eligible for review.
Results
RQ1: The few existing studies were recent and geographically diverse, primarily student-focused, mostly ChatGPT-focused, and primarily used cross-sectional surveys, but with different measurement scales.
RQ2: Most studies examined anthropomorphic response, while only one study examined anthropomorphic cues. Four studies linked higher anthropomorphism to positive effects on trust and learning, one found a negative link to trust, and three found no significant impact on student preferences and learning.
The review found mixed evidence but suggests that anthropomorphism can significantly influence trust, engagement, and learning. We highlight the need for standardised anthropomorphism assessments and (quasi-)experimental designs to establish causality, observe long-term effects, and rigorously assess interventions.
| Originalsprog | Engelsk |
|---|---|
| Publikationsdato | 23 aug. 2025 |
| Antal sider | 1 |
| Status | Udgivet - 23 aug. 2025 |
| Begivenhed | JURE 2025 - Graz University, Graz, Østrig Varighed: 23 aug. 2025 → 24 aug. 2025 https://www.earli.org/events/jure2025 |
Konference
| Konference | JURE 2025 |
|---|---|
| Lokation | Graz University |
| Land/Område | Østrig |
| By | Graz |
| Periode | 23/08/2025 → 24/08/2025 |
| Internetadresse |