{"resourceId":"berkeley-ai-assistance-reduces-persistence-colm-20261009","versions":[{"version":"external-a7573577d5900ce9ca7479223ed2a04df4fd53bf1c50a0b9c7af5949d4ff8e17","resource":{"id":"berkeley-ai-assistance-reduces-persistence-colm-20261009","title":"UC Berkeley: randomized trials (N=1,222) find about 10 minutes of AI assistance reduces persistence and unassisted accuracy once the tool is removed","organization":"UC Berkeley Center for Human-Compatible AI (Brian Christian) with Carnegie Mellon, MIT, Oxford and UCLA co-authors","sector":"Higher education student success / learning persistence and AI assistance (cross-cutting cognitive research)","geography":"Online participant samples (not campus-specific); research team in U.S. and U.K.","publishedAt":"Berkeley News 9 October 2026; peer-reviewed paper presented at the Conference on Language Modeling the same week (arXiv 2604.04721 v5, revised 3 Oct 2026)","publicationDate":"2026-10-09","eventDate":"2026-10-07","sourceName":"Berkeley News","sourceLabel":"University news article on a peer-reviewed conference paper reporting randomized controlled trials","sourceUrl":"https://news.berkeley.edu/2026/10/09/using-ai-for-just-10-minutes-erodes-your-ability-to-persist-at-hard-things/","evidenceClass":"academic-research","outcomeClass":"cautionary","topics":["knowledge-work","developers-agents","operating-model"],"finding":"Berkeley News reports randomized trials with 1,222 online participants. In the first experiment (354 participants) one group solved fraction problems unaided while another had ChatGPT available; the AI group started more accurate, but when the tool was removed after 12 problems its accuracy dropped almost immediately and skipping rose. A second experiment (667) replicated this, and a third (201) using an SAT reading-comprehension task showed the same persistence and accuracy drop. The updated peer-reviewed paper was presented at COLM this week. Christian argues AI responses could default to being more instructive, like a tutor.","sledRelevance":"Published today, this is causal evidence that answer-giving AI can erode the persistence that predicts long-term learning, which is central to student-success concerns about course AI defaults, tutoring design and assessment. Relevance to HE is indirect because participants were online adults on short tasks.","evidence":"Three RCTs, total N=1,222 (354, 667, 201); fraction and SAT reading tasks; solve and skip rates after AI removal; peer-reviewed COLM paper. Long-term effects and campus settings were not tested.","architectureImplications":"Interpretation: default chatbot behavior (instant complete answers) is the mechanism at issue; tutor-mode configurations that withhold answers or scaffold are the design response.","governanceImplications":"Interpretation: course AI policies and tool configurations should consider unaided assessment and persistence, not only assisted performance.","securityPrivacyImplications":"No student data involved; not applicable beyond normal research ethics.","caveats":"Online participants, not enrolled students in courses; short tasks of about 10 minutes; durability of effects unknown; news article summarizes rather than reports full statistics; a preliminary version circulated earlier in 2026.","streamIds":["student-success"],"roles":{"sales":"The framing point is that short-term help can cost long-term persistence. Berkeley-led randomized trials with 1,222 participants found that about 10 minutes of AI assistance boosted accuracy but, once the tool was removed, people answered worse and gave up more. Use it with academic leaders choosing AI tutor defaults. Keep it bounded: online adult participants on short tasks, not students in courses, and durability is untested.","engineering":"For engineering, the framing point is configuring answer behavior. The harm traced to instant complete answers, so evaluate tutor-mode settings that scaffold, hint or withhold answers, and log whether students attempt problems before asking. Proof of value should include unaided post-tests after assisted practice, because assisted accuracy alone hid the drop in these trials. Note the evidence comes from short online tasks rather than campus courses.","delivery":"Delivery's framing point is protecting productive struggle. In courses that permit AI, schedule periodic unaided checks, set default tool modes to hint before answering, and brief instructors and tutors on persistence risks. Define success as stable unaided performance, not just assisted scores. Treat the Berkeley effect as a risk to monitor locally, since campus settings and longer exposure were not studied."},"roleDetails":{"sales":{"customerProblem":"Campuses deploy answer-giving AI assistants that may raise assisted performance while weakening students' persistence and independent skill.","stakeholders":["Provost","Teaching and learning center","Academic technology","Tutoring and learning-support services","Gateway-course faculty"],"discoveryQuestions":["Do your AI tools default to full answers or scaffolded hints?","How do courses measure unaided performance when AI is allowed?","Who decides tutor-mode configurations?"],"valueHypothesis":"Scaffolded AI defaults and unaided checkpoints may reduce persistence loss; benefits in real courses are untested.","potentialEngagement":"A tutor-mode configuration review plus an unaided-assessment pilot in one gateway course.","unsupportedClaims":["That all AI use harms learning","Long-term or semester-scale effects","Effects for enrolled students specifically"]},"engineering":{"fit":"Campus AI tutor and chatbot deployments where answer behavior is configurable.","architectureIntegration":"Tutor-mode system prompts or vendor settings that hint first; LMS-linked unaided practice items to measure persistence.","prerequisites":["Configurable AI tool","Unaided assessment items","Analytics on attempts and skips"],"constraints":["Evidence from online adults on short tasks","Vendor defaults may not be adjustable","Measuring persistence requires new instrumentation"],"security":"Attempt and skip logs are education records; minimize and protect them.","proofOfValue":["Unaided post-test after assisted practice","Skip-rate comparison between answer and hint modes","Student satisfaction check to detect disengagement"]},"delivery":{"work":"Set hint-first defaults in campus AI tools, add unaided checkpoints in pilot courses, and monitor persistence indicators.","dependencies":["Academic technology","Pilot faculty","Vendor configuration support"],"ownership":"Academic technology owns configuration; faculty own course checkpoints; learning-support services own student coaching.","skillsAdoption":"Brief faculty and tutors on productive struggle and how to explain hint-first behavior to students.","governanceCheckpoints":["Configuration approval before term","Mid-term persistence review","End-of-term outcome review"],"acceptanceCriteria":["Unaided checkpoint scores do not decline relative to prior term","Skip rates on unaided items monitored and reported","Hint-first default enabled in all pilot sections"],"risks":["Student frustration with hint-first tools","Students switching to unrestricted outside tools","Overgeneralizing lab findings"]}},"retrievedAt":"2026-10-10T03:23:32Z","enrichedAt":"2026-10-10T03:27:23Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Suggests faculty and tutor training should address productive struggle when AI is available.","procurementImplications":"Ask vendors whether tutoring modes can withhold direct answers and scaffold, and for evidence on unaided performance.","operatingModelImplications":"Include unaided checks in courses where AI assistance is permitted.","sourceVerification":{"openedUrl":"https://news.berkeley.edu/2026/10/09/using-ai-for-just-10-minutes-erodes-your-ability-to-persist-at-hard-things/","referenceExcerpt":"spending just 10 minutes relying on an AI tool significantly impairs our ability to focus and persist with difficult tasks","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}