Actionable companion · Learning design & outcomes

Learning with agency: Physical embodiment and anthropomorphism of AI tutors and their role in student enjoyment and performance

AI-assisted editorial implementation guide, separate from the original research. Proposed actions are not claims or recommendations attributed to the source authors. No independent peer review of this guide is asserted.

Read original paper Search paper agent CC BY 4.0

Evidence and context

Title and abstract screening; stage labels describe reported settings, not verified numerical ages. Design labels require full-text appraisal. Reported stage(s): Age / stage not established. Verify the participant population and task before transferring the findings.

Original authors’ abstract, reproduced under the linked source license:

Abstract Rising interest in artificial intelligence in education reinforces the demand for evidence-based implementation. This study investigates how tutor agents’ physical embodiment and anthropomorphism (student-reported sociability, animacy, agency, and disturbance) relate to affective (on-task enjoyment) and cognitive (task performance) learning within an intelligent tutoring system (ITS). Data from 56 students (M = 17.75 years, SD = 2.63 years; 30.4% female), working with an emotionally-adaptive version of the ITS “Betty’s Brain”, were analyzed. The ITS’ agents were either depicted as on-screen robots (condition A) or as both on-screen avatars and physical robots (condition B). Physical presence of the tutor agent was not significantly related to task performance or anthropomorphism, but to higher initial on-task enjoyment. Student-reported disturbance was negatively related to initial on-task enjoyment, and student-reported sociability was negatively related to task performance. While physical robots may increase initial on-task enjoyment, students’ perception of certain characteristics may hinder learning, providing implications for designing social robots for education. Subject terms: Education, Psychology

What kind of evidence is this?

Intelligent tutoring & adaptive learning · Survey / observational study

Stage: Age / stage not established. Screening labels describe reported context and study-design language. They are not a risk-of-bias rating or proof of effectiveness for an age band.

Study methods · source passage

Originally, a total of 64 students from two secondary schools in Germany participated in the experiment. Due to incomplete data, n = 6 participants had to be excluded from analysis as they did not provide any data at the pre-test measurement, and n = 2 had to be excluded due to extremely outlying age values (≥25 years), resulting in a sample of N = 56 students (Mage = 17.75, SDage = 2.63; 30.4% female, 69.6% male). Participants were randomly assigned to one of the two experimental conditions. In condition A, supported only by a virtual robotic tutor agent, 51.8% of participants were included (ncon = 29 students: 41.4% female, 58.6% male). Meanwhile, condition B, in which students were additionally supported by a physically present robotic tutor agent, comprised 48.2% of participants (nexp = 27 students: 18.5% female, 81.5% male). The majority of the participants were native German speakers (78.6%). Prior to participation, informed consent was obtained from all participants and, if required, their parents or legal guardians. The study was performed in line with the principles of the Declaration of Helsinki. It received approval from the University of Potsdam Ethics Committee (number 78/2021) as well as the Ministry of Education, Youth and Sports of the Federal State of Brandenburg (number 22/2022).

Limitations and uncertainty · source passage

This study investigated the role of the physical embodiment of a pedagogical tutor agent of an AI-based adaptive system in addition to its virtual avatar in emotional and cognitive learning outcomes and students’ perceptions of the tutor agent’s anthropomorphic characteristics. In addition, the relevance of anthropomorphic characteristics for students’ emotional and cognitive learning was investigated. Key findings indicate that the physical (non-)presence of the pedagogical agent was not significantly associated with either (1) student-reported anthropomorphism or (2) task performance. However, (3) when physically present, the robotic tutors were significantly related to higher on-task enjoyment in the first half of the interaction (novelty phase) in two of the four models (Model 1Soc and Model 2Anim). Further, (4) a higher student-reported perception of sociability was associated with lower task performance, and (5) a higher student-reported perception of disturbance was related to lower on-task enjoyment during the novelty phase. In the following sections, we discuss these findings along with our hypotheses and specify potential implications for the implementation of social robots as AI-based tutors in learning and education, while acknowledging the limitations of our study, including the small sample size and specific experimental setup.

Design the system before the pilot

Document the learner model, mastery rule, hint sequence and adaptation signal. Compare adaptive routing with the same content in a non-adaptive sequence. Check whether early errors lead to persistent low-expectation pathways.

Specify a measurable learning objective, a comparator and a task without tool assistance before adopting a system.

A 90-day district pilot

Days 1–15 · Appraise. Have the curriculum lead and a teacher read the methods and limitations. Record the study design, sample, comparator, outcomes and evidence gaps. Define a local question that the study can help inform.
Days 16–30 · Design. Treat the paper as background evidence until its population is verified in the full text. Do not assign an age-specific intervention or extend adult results to children. Name an accountable owner and set the baseline, success threshold and budget before implementation.
Days 31–60 · Test. Start with a small voluntary cohort, preserve an accessible non-AI route and require teacher review. Use synthetic or de-identified examples; approve data handling and tool terms through district processes.
Days 61–90 · Decide. Extract the participant ages or education stage, recruitment, task and comparator before selecting a local pilot. If these cannot be established, retain the paper for design discussion rather than age-band efficacy claims. Compare with baseline, document uncertainty and stop or revise if learning worsens, access is unequal or critical errors persist.

The board decision brief

Report the learning problem, evidence quality, participating population, total cost including review time, measured benefit, errors, access differences and remaining uncertainties. Make a dated recommendation to stop, revise or expand. A positive perception survey alone does not establish learning improvement.

Attribution and reuse

Helene Ackermann; Anja Henke; Johann Chevalère; Hae Seon Yun; Verena V Hafner; Niels Pinkwart; Rebecca Lazarides. Physical embodiment and anthropomorphism of AI tutors and their role in student enjoyment and performance. NPJ Science of Learning (2025-01-08). https://doi.org/10.1038/s41539-024-00293-z

Companion prepared 2026-10-30. Original authorship and DOI are preserved. The abstract is an attributed source extract; the implementation plan is new AI-assisted editorial material, offered under the same CC BY 4.0 license. This guide has not undergone independent peer review.