AIEd Knowledge BaseAIEd KB
On this page

Concept

Trust

created 2026-08-13 · updated 2026-11-10

Trust — the willingness of learners, educators, and institutions to rely on a person or an AI system for learning, judgment, and decision-making. In AI in education, trust spans two related but distinct domains: trust in AI (confidence in the competence, transparency, reliability, and benevolence of an AI system or agent) and interpersonal trust (the relational trust between students and instructors, between learners and peers, and across the institution). Both are double-edged: appropriate trust enables productive engagement, while over-trust invites over-reliance and under-trust blocks beneficial use. The central challenge is calibration — aligning trust to actual reliability, whether that reliability belongs to a model or to a person.

Questions to Consider

  • When you say you 'trust' an AI tool versus trusting a teacher or a colleague, are you describing the same thing? What is similar, and what is fundamentally different, about placing trust in a system versus in a person?
  • There's a documented 'trust-utility gap': a tool's apparent competence often exceeds its actual reliability. Recall a tool that looked impressive but let you down, or one that seemed limited but proved dependable. What shaped the gap between how it looked and what it could really do?
  • Consider an AI that always agrees with you and never challenges your ideas. It might feel comfortable and trustworthy — but is agreement the same as reliability? What might you be giving up if the tool you rely on never pushes back?
  • The page argues that an instructor's trust in AI shapes how students trust that AI — the two domains interact. In a course you know well, how would a teacher's enthusiasm or skepticism about AI influence whether students accepted or questioned the tool?
  • Students often decide whether to disclose their AI use based on comfort with their instructor more than on policy. If you were (or are) a student, what would make you willing to be honest about using AI — and what would make you hide it? What does that say about how trust is actually built in a classroom?
  • One finding: an AI tutor that warns 'I may make mistakes' prompted students to seek more help, not less. What does this suggest about whether acknowledging limits undermines or strengthens the trust that supports real learning?

Introduction

Trust in AI is shaped by perceived competence, transparency, consistency, and whether the system appears aligned with the learner's goals; it is closely tied to AI Literacy (knowing what to trust), Critical Thinking (evaluating output), and the design of responsible AI. Interpersonal trust, by contrast, is built through relationships, disclosure, feedback, and pedagogical care — the qualities students rely on when they decide whether an instructor or an AI is a trustworthy source of guidance. The two domains increasingly interact: AI is woven into teacher-student relationships, so how students trust their instructor shapes how they trust (or question) the AI tools that instructor endorses.

Trust is a measured casualty, not only a principle. In Gillette et al. (2026), students said they trusted classmates less after discovering undisclosed genAI use in group work, calling it "detrimental to the trust and ethics of academia", while a few suspected teachers used genAI and withheld that from students.

Trust in AI systems

Research in this knowledge base examines when learners appropriately trust AI-generated guidance. Warnings about AI fallibility can improve calibration: a simple transparency intervention telling students an AI tutor may make mistakes increased Help-Seeking in a math ITS, suggesting that honest limits foster rather than undermine appropriate reliance. Co-designing trustworthiness metrics with learning engineers shows that trust is best built on observable, agreed-upon criteria rather than assumed capability. Physics and image-generation studies reveal a persistent trust-utility gap — users must weigh a tool's apparent competence against its actual reliability in a task. Among the youngest users, Vahedian Movahed & Martin (2025) found that 52% of children (ages 6–14) generally trusted an age-tailored chatbot and 35% trusted it like a teacher or friend, with about a third willing to confide in it; children also actively tested its credibility with known-answer questions, and trust showed no statistically significant grade-level differences — illustrating how early trust can form ahead of critical evaluation.

Transparency about what a system is — not only what it can do — also shapes calibration: the Amico accompaniment prototype made its identity and limits explicit, and its Italy–China pilot found learners recognized it as a bounded support tool rather than an autonomous tutor or relational substitute, avoiding anthropomorphic over-trust (Benedetti (2026)).

In higher-education institutions, the lever may be usefulness rather than rules: among 2,121 students, faculty and staff surveyed at a single university, perceived usefulness outweighed policy clarity as a correlate of AI trust (β = 0.402 against 0.223) (Braun & Khafizov, 2026). Distrust of accuracy can coexist with confidence in managing the tool. In a survey of 8,021 students across four Australian universities, 51% distrusted GenAI's factual accuracy while 76% were confident in their ability to get the outputs they wanted, and usefulness rather than trust or rule compliance drove their use (Chung et al., 2026).

The modeling of AI overreliance as a complex adaptive system reframes trust as a population-level process: whether people trust an assistant when it is right and check it when it is wrong depends on social dynamics and feedback loops, not just individual judgment. Sycophancy threatens calibration from the other direction — an AI that always agrees can feel trustworthy precisely because it never challenges the user, inviting uncritical acceptance (contextual sycophancy and sycophantic AI in social interaction). In embodied contexts like Robots in Education, trust is shaped more by what the robot does than what it looks like (task context and trust in educational HRI), and avatar identity shapes the epistemic trust learners place in AI content. Trust in analytics tools is also context-dependent. Mejia-Domenzain et al. (2026) found that teachers' concerns and adoption barriers diverged sharply by learning context: flipped-classroom (university) teachers worried most about data anonymization and student opt-out, whereas reflective-writing (vocational) teachers feared misuse of the tool by fellow educators and stressed the need to contextualize data — even though both groups reported similar Self-Efficacy and perceived benefits in a trust in AI survey. The finding that trust in the tool is decoupled from trust in its data governance and social use underscores that building appropriate trust in analytics requires attending to context-specific concerns, not just the system's apparent competence.

How explainable a system is — and in what terms — also shapes whether teachers trust its recommendations. In a within-subject experiment with 41 in-service chemistry teachers using the AI grouping-recommendation tool GrouPer, Feldman-Maggor et al. (2025) found that explainable AI builds trust indirectly by increasing the understandability of the system's performance, and that domain-driven explanations framed in curricular/pedagogical language fostered significantly greater understandability and learned trust than purely data-driven (feature-importance) explanations. Notably, understandability alone was insufficient for some teachers — they reported needing real classroom experience with the tool before fully relying on it — reinforcing that trust in AI is dynamic and validated through situated use, not granted by explanation alone.

Risk perception and trust are not opposites. A 130-student perception study of agentic GenAI in higher education found perceived risk moderately elevated (M = 3.33, SD 0.89) alongside more favorable trust and adoption intention (M = 3.62, SD 0.81), and — contrary to a simple deterrence expectation — a positive association between perceived risk and continued-use intention (Spearman's ρ = 0.317, p < 0.001) (Ilieva et al. 2026). The authors read this as informed adoption rather than indifference: engaged or experienced users recognize both the value and the limits of the technology, and only 45.4% said they trusted agents under instructor guidance. For Trust Calibration, the implication is that awareness of risk is not the absence of trust — it can be a component of it — while the cross-sectional design leaves awareness, exposure, and self-selection indistinguishable.

Trust in AI behaves as a pathway from appraisal to reliance rather than a single attitude: perceived AI empowerment and perceived AI threat were near-independent (r=0.030) yet moved trust in opposite directions (β=0.576 and β=−0.328), and trust carried significant indirect associations with GenAI dependence in both directions (Zhu et al. (2026)).

A structural model gives that bridging role an empirical test. In Bu & Li (2026)'s survey of 642 university students, contextual support predicted trust calibration (β = 0.56), trust calibration predicted collaborative AI Regulation in Education (β = 0.49) and perceived transfer gains (β = 0.22), and the direct path from contextual support to gains was not significant.

Trust ahead of every classic acceptance path. Zhu (2026) added trust to UTAUT in a survey of 450 art-and-design students in Zhejiang and found it the strongest direct predictor of intention to use generative AI (β = 0.425). Social influence (0.155), metacognitive control (0.146) and performance expectancy (0.128) followed it. Social influence also carried the largest path into trust (β = 0.262), and 42.1% of its total association with intention ran through trust.

The construct is thinner than the coefficient implies. It measures instrumental technology trust with an art-value supplement — stated confidence rather than calibrated reliability — and the two authorship items were dropped after forming their own factor in a 121-student pilot. A student can trust this model appropriately or indiscriminately, and the instrument cannot tell the two apart.

Transparency can be the strongest way into trust without trust being the strongest way into use. Wang et al. (2026) surveyed 653 learners about AI avatars and found transparency predicted trust at β = 0.542 — the largest path in their model — while that trust predicted adoption intention at only β = 0.196, behind relative advantage (0.267) and perceived usefulness (0.247).

AlGhamdi (2026) supports a function-specific rather than global account of trust in algorithms: neither algorithm aversion nor algorithm appreciation described these 13 students, who simultaneously trusted ChatGPT for surface-level feedback and distrusted it as a grader. Their trust was conditional on instructor oversight, and their skepticism stemmed from the AI's contextual limits — misreading scanned handwriting ("it spelled my last name wrong and it thinks I made mistakes in my spelling when I didn't"), not knowing the instructor's rating system, and chronic positivity — rather than from technophobia, suggesting trust measures should be decomposed by the function an AI performs and the stakes attached to it.

The three Source Credibility dimensions can dissociate: seventeen EFL undergraduates granted an AI writing assistant broad expertise, doubted its trustworthiness mainly over what the platform retained, and rated its goodwill lowest - the ceiling on their reliance sat in care, not competence (Serpil & Mor, 2026).

Credibility cues can raise reliance on wrong answers: adding citations and authority signals to a debate interface raised abandonment of a correct judgment for the AI's position from 15.0% to 31.8%, while withholding a recommended position cut that switching to 5.7–8.2% (Tuo et al., 2026).

Trust in AI does not track national uptake either: Romanians, at 17.8% generative-AI use against Denmark's 48.4%, rated their trust in AI slightly higher (4.1 against 3.8 on a seven-point scale) while institutional trust split 28 to 45 points the other way (Lucau, Voicu & Christensen (2026)).

Trust in AI tracks psychological state, not demographic category

Kumar et al. (2026) clustered 107 students at a public HBCU on resilience, perceived stress and AI trust, and found three profiles in which trust in AI dissociated from confidence in oneself: a high-resilience low-stress group with favorable AI trust, a moderately stressed group holding the highest AI trust of the three despite strain, and a psychologically resilient group that was nearly as resilient as the adopters but markedly lower in AI trust. Stress and AI trust separated the clusters most strongly (partial eta squared 0.528 and 0.521 against 0.315 for resilience), and gender was the only demographic variable significantly associated with membership, with STEM affiliation, academic level, employment status and age group all non-significant. Two implications matter for this page: first, low trust in AI is not a proxy for low confidence or low technical familiarity, since the skeptics were the most resilient and the most STEM-heavy group, which the authors read as calibrated skepticism rather than resistance; second, because the clusters were only weakly separated (silhouette 0.288, with the Calinski-Harabasz index preferring two clusters) they should be treated as overlapping profiles rather than distinct student types. The findings come from one institution and a cross-sectional self-report survey, so they establish that trust varies with psychological state, not why.

Interpersonal trust in education

Trust is also fundamentally relational. The classroom trust gap is documented in teacher-student views on control and agency in K-12 AI: students want greater autonomy and flexibility while teachers prioritize oversight and monitoring, a misalignment that both sides must navigate for AI adoption to succeed. Why students disclose or conceal their AI use shows that disclosure is driven less by policy than by relational factors — perceived peer norms and comfort with instructors are the strongest predictors, pointing to low interpretive trust in institutional settings. The hidden costs of disclosure add nuance about when honesty is socially costly.

A detection-first response works the other way: unreliable detectors that disproportionately flag nonnative speakers stall misconduct cases for lack of proof, producing "suspicion without recourse", whereas five-to-ten-minute verification conversations announced as pedagogy rather than surveillance restored faculty authority and drew more open AI disclosure (Mandernach, 2026).

Reliance on detectors and surveillance, Bassett et al. (2026) argue, fosters a climate of suspicion that erodes student trust in assessment and the institution — the opposite of the trust integrity processes are meant to protect.

Feedback is a key site of interpersonal trust. Students' perceptions of GenAI versus teacher feedback find the two serve different needs — complementary but not interchangeable — with students trusting teacher feedback for relational, personalized judgment and GenAI for speed and Accessibility. A "care-full" account of feedback argues that trustworthy feedback is an ethical, relational practice: it builds educative relationships and is respected as a professional craft, values an AI cannot simply replicate. This is why teacher-student trust — built on care and professional judgment — remains central even as AI enters the feedback loop.

Peer trust runs through a different channel: among 406 university students rating a recalled classmate, that classmate's GenAI use was associated with lower interpersonal trust, carried by perceived warmth and competence, and the negative association held only when perceived AI literacy was low — it was non-significant at high (Zhang et al. (2026)).

Calibration and the two domains together

The unifying challenge is calibration: matching trust to actual reliability, whether the trusted party is a model or a person. Trust Calibration is the metacognitive capacity to know when to trust and when to question. Studies of AI Feedback and Intelligent Tutoring examine when learners appropriately rely on or challenge AI guidance, while the interpersonal literature shows that students' trust in an instructor depends on relational trust built over time. As AI becomes embedded in teaching, these domains converge: an instructor who transparently explains what an AI tool can and cannot do, and who demonstrates reliability in their own judgment, builds the kind of trust that carries over to the tools they endorse. Building appropriate trust — in both AI and in each other — is a core goal of responsible AI design in education.

Calibration is also tested by the incentives of the trusted system itself. When staff skeptical of AI adoption consult conversational AI — built by organizations with a commercial stake in adoption — there is a risk the system is predisposed to encourage it. An audit of ten frontier models found most acknowledged a rural K-12 staff member's concerns (job threat, being 'not for people like me') before redirecting toward engagement. This challenges naive reliance on trust and underscores the importance of human oversight and independent evaluation of AI advice.

Common Misconceptions

"Trust and distrust sit at opposite ends of one scale." They are separable, and confusing them misreads what students actually report. Risk perception and trust are not opposites: a 130-student perception study of agentic GenAI in higher education found perceived risk moderately elevated (M = 3.33, SD 0.89) alongside more favorable trust and adoption. A learner can hold both at once, which is why a single trust score cannot tell you whether their reliance is appropriate.

"Distrust of AI's accuracy means students won't lean on it." Students distinguish the output from their ability to steer it. In a survey of 8,021 students across four Australian universities, 51% distrusted GenAI's factual accuracy while 76% were confident in their ability to get the outputs they wanted. The second number is the one that predicts behavior: confidence in extraction coexists with doubt about truth, and only one of those is a safety property.

"Clearer rules and policies are what build student trust." Usefulness dominates. A survey of 2,121 students, faculty and staff at one university found perceived usefulness the strongest correlate of trust in AI (β = 0.402), with perceived policy clarity trailing it. Institutions that invest in rule-making while shipping tools nobody finds useful have aimed at the weaker lever.

"If learners trust it more, they will use it more carefully — or at least appropriately." The opposite is the documented pattern. Among 432 undergraduates solving Python problems with accurate and deliberately misleading suggestions, greater trust predicted lower appropriate reliance (r = −.42), and students accepted 86.03% of the misleading suggestions. Trust is not a proxy for calibration; see trust calibration.

"Honest warnings about fallibility would undermine confidence in a tutor." They improve behavior. Warnings about AI fallibility in a math ITS increased Help-Seeking rather than reducing use, which suggests that stating limits fosters appropriate reliance instead of eroding it.

Implications for Practice

For instructors

  • Measure two things, not one: what students believe about AI's accuracy, and what they believe about their own ability to use it. The 51%/76% split shows these diverge sharply and only the second drives behavior.
  • State the tool's limits up front. Fallibility warnings raised help-seeking in the ITS evidence, so disclosure is a calibration tool rather than a confidence risk.
  • Watch for high trust paired with shallow checking; that combination is the calibrational failure mode, not a success signal.

For developers

  • Build trust on observable, agreed criteria rather than on apparent capability, following the co-design-of-trustworthiness work with learning engineers. Fluency is not a trustworthiness signal and students read it as one.
  • Prioritize usefulness in onboarding over policy copy, since usefulness is the strongest trust correlate measured.
  • Do not treat high trust as engagement success; pair any trust metric with an appropriate-reliance measure.

For researchers

  • Report trust, risk perception and appropriate reliance as distinct constructs, because they co-vary in non-obvious directions — including negatively.
  • Report the behavioral outcome (acceptance rate of wrong suggestions) alongside self-report trust; trust scales alone invert the conclusion.

Connected Concepts

Connected Articles

Metadata

Technologies and techniquesRobots in Education, Intelligent Tutoring
Equity, ethics, and responsible useTrust
Foundations of AI in educationAI Literacy, Critical Thinking, Human AI Collaboration
SourceHistoryEditChat

Embed this page

Copy the code below to embed this page in a learning management system or other website. The embedded version shows only the page content — it hides the site header, navigation, and footer.