[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-ai-tools-cut-study-time-on-text-math-problems-hurt-retention":10},{"siteName":4,"siteTagline":5,"publisherName":4,"contactEmail":6},"The Revision","Tech news, decoded.","editor@therevision.news",{"gaMeasurementId":8,"adsenseClientId":9},"G-ZW2MV82GYR","ca-pub-8533917693782264",{"article":11},{"id":12,"slug":13,"title":14,"dek":15,"body_md":16,"tags_json":17,"published_at":18,"created_at":19,"updated_at":20,"status":21,"review_note":22,"review_notes":23,"image_url":22,"persona_id":22,"persona_name":22,"section":22,"tags":24,"sources":28,"feedback":32,"feedback_at":22,"cost_usd":32,"total_tokens":32},1397,"ai-tools-cut-study-time-on-text-math-problems-hurt-retention","AI tools cut study time on text math problems, hurt retention","A decade‑long analysis of 3.2 million ALEKS interactions shows students spend less time on AI‑friendly word problems and score lower on unproctored tests.","- Students are spending noticeably less time on math word problems that can be answered by generative AI.\n\nA ten‑year panel of 3.2 million ALEKS learning interactions reveals a 2.8% quarterly drop in time‑on‑task for college students after ChatGPT’s release, adding up to a 26.9% decline over eleven quarters. High‑schoolers cut time by 31.3%, middle‑schoolers by 9.0%, while Grade 5 shows no change. The study isolates AI‑susceptible text problems from interactive graph tasks, using a quasi‑experimental design. In proctored settings the time reduction disappears, suggesting the drop reflects AI use rather than faster learning. Logistic fixed‑effects models show a 25% cumulative drop in odds of a correct answer on proctored retention items, while unproctored assessments show the opposite trend.\n\nThe findings hint at a broader “cognitive surrender”: students lean on AI for easy‑type problems, reducing practice and retaining less knowledge. This pattern raises questions for educators about assessment design and for policymakers about the limits of AI‑assisted learning.\n\nIf institutions tighten proctoring, the AI‑driven efficiency gain evaporates—leaving the core issue of how to preserve deep learning in an AI‑pervasive world.","[\"education\",\"generative-ai\",\"assessment\"]","2026-06-16T04:00:00.000Z","2026-06-17T07:40:46.739Z","2026-06-17T07:40:49.555Z","published",null,[],[25,26,27],"education","generative-ai","assessment",[29],{"name":30,"url":31},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2605.21629",0]