{"resourceId":"lindqvist-quince-engineering-students-genai-survey-20261012","versions":[{"version":"external-79c82baebda629970e8d5b46659deaf15709f2e7d0b990330f3010ed5c808147","resource":{"id":"lindqvist-quince-engineering-students-genai-survey-20261012","title":"Peer-reviewed survey of 131 Australian engineering students finds GenAI use varies sharply by task and cohort, with accuracy verification the top challenge","organization":"University of Technology Sydney and Southern Cross University (authors Lidfors Lindqvist and Quince)","sector":"Higher education student success / AI literacy and student GenAI use in engineering","geography":"Australia (two universities, metropolitan and regional); transferability to U.S. institutions limited","publishedAt":"STEM Education 6(6), published 12 October 2026","publicationDate":"2026-10-12","eventDate":null,"sourceName":"STEM Education (AIMS Press)","sourceLabel":"Peer-reviewed exploratory survey study","sourceUrl":"https://www.aimspress.com/article/doi/10.3934/steme.2026042","evidenceClass":"academic-research","outcomeClass":"mixed","topics":["knowledge-work","accessibility-workforce","governance-procurement"],"finding":"A peer-reviewed survey of 131 engineering students at two Australian universities found GenAI use was broad but uneven: 35.71% of a 28-student first-year regional cohort reported no academic use versus 9.71% of a 103-student second-year metropolitan cohort, and code generation was reported by 7.14% versus 46.60%. Verifying accuracy was the most common challenge for users in both cohorts (66.67% and 77.42%), and over 95% agreed universities have a role in preparing them for AI. The authors recommend gradual, discipline-specific GenAI literacy integrated into coursework.","sledRelevance":"Published 12 October, this adds fresh peer-reviewed evidence that student AI use is task- and stage-dependent, which bears directly on whether generic AI literacy courses or embedded discipline-specific instruction will meet student needs.","evidence":"Cross-sectional survey, 11 questions, chi-square and Mann-Whitney tests with Bonferroni adjustment and 95% CIs; cohorts are about 30% of each class. Self-reported use and perceptions only.","architectureImplications":"Limited relevance; GPT-based tools dominated use in both cohorts, with Google and Bing AI usage differing by institution.","governanceImplications":"Policy awareness similar across cohorts; perceived training on AI policy differed modestly (means 4.25 vs 3.78). Authors argue for literacy aligned to discipline tasks rather than compliance-first policy.","securityPrivacyImplications":"Not addressed in the study.","caveats":"Exploratory; confounded cohorts; self-report; engineering at two Australian universities only; the study does not measure learning, competence or misconduct.","streamIds":["student-success"],"roles":{"sales":"The framing point is that student AI use is task- and stage-specific, so one-size literacy offerings may miss. This peer-reviewed survey of 131 Australian engineering students found code generation reported by 46.60% of a second-year cohort versus 7.14% of a first-year cohort, and accuracy verification was the top challenge. That supports conversations about discipline-embedded literacy. Keep claims bounded: it is a self-reported, cross-sectional study with confounded cohorts.","engineering":"For engineering, the framing point is that verification, not access, is the reported skill gap. Users in both cohorts named verifying accuracy as their main challenge, and coding use rose with curricular exposure. Pre-sales work on engineering course tools should prioritise test harnesses, code explanation and provenance features over general chat. The study did not measure competence, so proof of value must come from assessed verification tasks.","delivery":"Delivery's framing point is building verification routines into discipline courses year by year. The authors recommend gradual, discipline-specific GenAI literacy integrated into coursework with clear ethical guidance. Delivery teams would need course-level mapping of permitted AI uses, embedded verification exercises and a short competence check rather than awareness surveys. Proposed acceptance criteria are targets, not observed results, transferable mainly to engineering and computing programs."},"roleDetails":{"sales":{"customerProblem":"Engineering programs see uneven, task-dependent student AI use and weak verification skills, while generic policies and training may not match course tasks.","stakeholders":["Engineering faculty and associate deans","Teaching and learning centre","Academic integrity office","Student representatives"],"discoveryQuestions":["Which course tasks drive student AI use in each year of the program?","How are students taught to verify AI-generated code and calculations?","Does AI-policy training test comprehension or just awareness?","Do regional or older students need different support?"],"valueHypothesis":"Interpretation: discipline-embedded verification instruction may address the challenge students report most, more directly than generic literacy modules.","potentialEngagement":"A bounded curriculum-mapping exercise of AI use and verification tasks across one engineering program's years.","unsupportedClaims":["That AI use improves learning or competence (not measured)","That one institution's policy caused cohort differences","Generalisation beyond engineering at two Australian universities"]},"engineering":{"fit":"Fit is for course tools that support verification, such as testing, code explanation and source checking, within engineering coursework.","architectureIntegration":"Interpretation: integrate verification exercises and AI tools into existing LMS and coding environments.","prerequisites":["Course-level mapping of permitted AI uses","Access to sanctioned AI tools for students"],"constraints":["Self-report evidence only","Confounded cohort comparison","Discipline-specific findings"],"security":"Not addressed; any sanctioned tool should meet institutional data terms for student code and work.","proofOfValue":["Assess students on detecting errors in AI-generated code","Compare verification performance before and after embedded instruction"]},"delivery":{"work":"Embed staged GenAI literacy and verification exercises into engineering courses across years, with a short competence check.","dependencies":["Program curriculum committee","Course coordinators","Teaching centre","Integrity office guidance"],"ownership":"Engineering school teaching and learning leadership.","skillsAdoption":"Instructors need examples of verification tasks; students need explicit rules distinguishing permitted support from independent work.","governanceCheckpoints":["Curriculum committee approval of staged outcomes","Integrity office review of course AI rules","Annual review of competence check results"],"acceptanceCriteria":["Proposed: each year of the program includes at least one assessed AI-verification task","Proposed: competence check shows improvement between first and second year","Proposed: course AI-use rules published in all core courses"],"risks":["Added tasks crowd existing content","Self-report surveys mistaken for competence evidence","Uneven adoption across courses"]}},"retrievedAt":"2026-10-14T03:25:07Z","enrichedAt":"2026-10-14T03:34:21Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Regional first-year cohort was older and more often non-users; interpretation: literacy support may need to account for older and regional learners.","procurementImplications":"None directly; interpretation: tool choices should follow task-level needs such as coding support.","operatingModelImplications":"Supports embedding verification routines and AI guidance in discipline courses across years of study.","sourceVerification":{"openedUrl":"https://www.aimspress.com/article/doi/10.3934/steme.2026042","referenceExcerpt":"Our findings suggest that GenAI literacy development should be gradual, discipline-specific, integrated into coursework, and guided by clear communication about the ethical use of GenAI.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}