How Students With Motor Disabilities View AI Through Lens of Academic Confidence

The researchers recommend giving students AI tools that are easy to use, along with guidance that builds their confidence and helps them check the answers they receive. A writing assistant or study chatbot is less useful if students struggle to operate it or do not understand how it can help them.

Getting through university involves more than understanding lectures and meeting deadlines, especially for students with motor disabilities who may face inaccessible spaces, digital barriers and limited support. A study published in Frontiers in Education, titled 'Between perceived competence and artificial intelligence: academic self-efficacy profiles in university students with motor disabilities,' explores how students' confidence in managing academic demands relates to the support they associate with AI.

Raquel Suriá-Martínez and colleagues found that students with stronger academic confidence reported higher ratings of AI across learning, information and emotional support. The findings raise a practical question for universities: how can access to these tools become meaningful for students with different levels of confidence and different support needs?

The confidence behind everyday university tasks

The researchers studied 102 students with motor disabilities at the University of Alicante and Miguel Hernández University of Elche in Spain. Participants were aged 18 to 33, with 54 women and 48 men in the sample, and were recruited with help from university disability support services.

A 13-item questionnaire examined three parts of academic confidence. Attention covered concentrating on tasks and processing information; Communication concerned exchanging ideas and interacting in learning settings; Excellence reflected planning, setting goals and working towards academic achievement.

Students completed a separate 12-item questionnaire about AI's perceived usefulness for educational, informational and emotional support. An independent group of 85 students with motor disabilities provided preliminary evidence supporting the questionnaire's three-part structure.

The researchers grouped participants according to their confidence scores, identifying 30 students with low self-efficacy, 42 with moderate self-efficacy and 30 with high self-efficacy. These groups represented differences in reported confidence across all three dimensions, without establishing fixed labels for students' ability or potential.

Where students saw the most value in AI

Educational support included understanding course content, organising study, generating ideas and working through academic problems. On the five-point questionnaire, average scores rose from 2.41 in the low-confidence group to 3.56 in the moderate group and 4.68 in the high-confidence group, producing the largest difference across the three support categories.

Informational support covered finding, organising and making sense of academic information, including explanations that could complement traditional study materials. Average ratings were 2.83, 3.74 and 4.51 for the low-, moderate- and high-confidence groups respectively.

Emotional support involved encouragement, guidance through academic difficulties, coping suggestions and a sense of companionship during study. Scores followed the same pattern, rising from 2.76 to 3.31 and 3.96, with a smaller gap between groups than for educational or informational support.

The authors frequently describe these findings as differences in AI use, although the questionnaire measured students' agreement with statements about AI's usefulness. These ratings capture students' perceptions and should not be read as direct records of how often they opened an AI tool or how effectively they used it.

A separate statistical model found positive links between all three confidence dimensions and the AI outcome, with Excellence showing the strongest association, followed by Communication and Attention. The model accounted for 43% of the variation in the measured outcome within this sample.

Planning, goal-setting and confidence about academic achievement stood out as particularly relevant to students' relationship with AI. The findings do not establish that these qualities caused greater engagement with the technology.

What useful support could look like on campus

The researchers recommend giving students AI tools that are easy to use, along with guidance that builds their confidence and helps them check the answers they receive. Students who feel less confident about their studies could start with small, manageable tasks, watch others use the tools and receive helpful feedback on their progress. More confident students could support their classmates or explore advanced uses of AI, with guidance suited to their needs.

Teachers need training in using AI and making lessons accessible, including offering students different ways to learn and show what they understand. Learning to check sources, spot bias and recognise incorrect or incomplete AI answers is an essential part of this support. Universities need to protect students' personal information and help them avoid becoming too dependent on these tools.

AI should be part of a wider effort to help students participate in university life and study independently. Accessible buildings, easy-to-use websites and reliable disability support services remain essential. An AI tool might offer encouragement during a difficult assignment, but it cannot replace the care and guidance provided by teachers, mentors, university staff or mental health professionals.

What this study can tell us and what needs testing

The study offers a snapshot of confidence and AI-related perceptions at one point in time. It cannot determine whether confidence encourages AI use, whether experiences with AI influence confidence, or whether the relationship works in both directions.

Its small convenience sample from two universities limits how widely the findings can be applied, including to students with other disabilities or the broader student population. Self-reported answers may overestimate engagement, and the new AI questionnaire requires further validation in larger independent samples.

The statistical model did not include device access, previous digital training, platform accessibility or campus support, leaving important possible explanations for differences between students unexamined. Higher ratings of AI's usefulness do not demonstrate better grades, stronger learning or improved wellbeing.

The authors call for studies that follow students over time, test confidence-building programmes and combine questionnaires with usage records or practical tasks. Research involving more diverse students and different AI tools could help clarify which forms of support are useful and under what conditions.