{"paper":{"title":"Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"Relying on LLMs for essay writing reduces brain connectivity and builds cognitive debt compared to writing unaided or with search tools.","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Ashly Vivian Beresnitzky, Eugene Hauptmann, Iris Braunstein, Jessica Situ, Nataliya Kosmyna, Pattie Maes, Xian-Hao Liao, Ye Tong Yuan","submitted_at":"2025-06-10T15:04:28Z","abstract_excerpt":"This study explores the neural and behavioral consequences of LLM-assisted essay writing. Participants were divided into three groups: LLM, Search Engine, and Brain-only (no tools). Each completed three sessions under the same condition. In a fourth session, LLM users were reassigned to Brain-only group (LLM-to-Brain), and Brain-only users were reassigned to LLM condition (Brain-to-LLM). A total of 54 participants took part in Sessions 1-3, with 18 completing session 4. We used electroencephalography (EEG) to assess cognitive load during essay writing, and analyzed essays using NLP, as well as"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Brain-only participants exhibited the strongest, most distributed networks; Search Engine users showed moderate engagement; and LLM users displayed the weakest connectivity. Over four months, LLM users consistently underperformed at neural, linguistic, and behavioral levels.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That observed EEG connectivity differences reflect accumulation of cognitive debt rather than short-term task demands or group differences, and that the small session-4 subsample (18 of 54) supports claims of long-term effects.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"LLM use for essay writing correlates with reduced brain network connectivity, lower self-reported ownership, and poorer recall of one's own content compared to unaided or search-based writing.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Relying on LLMs for essay writing reduces brain connectivity and builds cognitive debt compared to writing unaided or with search tools.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"9e25dd5af145eee5e47527c57dfe6dadaaffbf36639890f7372f5eb37a363b38"},"source":{"id":"2506.08872","kind":"arxiv","version":2},"verdict":{"id":"71094076-edba-4cd3-8405-ee5f0dbfe1e3","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T04:43:09.105299Z","strongest_claim":"Brain-only participants exhibited the strongest, most distributed networks; Search Engine users showed moderate engagement; and LLM users displayed the weakest connectivity. Over four months, LLM users consistently underperformed at neural, linguistic, and behavioral levels.","one_line_summary":"LLM use for essay writing correlates with reduced brain network connectivity, lower self-reported ownership, and poorer recall of one's own content compared to unaided or search-based writing.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That observed EEG connectivity differences reflect accumulation of cognitive debt rather than short-term task demands or group differences, and that the small session-4 subsample (18 of 54) supports claims of long-term effects.","pith_extraction_headline":"Relying on LLMs for essay writing reduces brain connectivity and builds cognitive debt compared to writing unaided or with search tools."},"references":{"count":144,"sample":[{"doi":"10.3389/feduc.2024.1392091","year":2024,"title":"Peláez-Sánchez, I. C., Velarde-Camaqui, D., & Glasserman-Morales, L. D. (2024). The impact of large language models on higher education: Exploring the connection between AI and Education 4.0. Frontier","work_id":"164d9d82-05b8-4eb4-ad89-0d362579a9de","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.3390/soc15010006","year":2025,"title":"one-person unicorn","work_id":"7a59461f-1337-4878-962f-3659eefb4f89","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.3389/fcogn.2023.1203077","year":2023,"title":"Shanmugasundaram, M., & Tamilarasu, A. (2023). The impact of digital technology, social media, and artificial intelligence on cognitive functions: A review. Frontiers in Cognition, 2, 1203077. https:/","work_id":"f98ceeec-3062-4e42-a135-6e8550b41f60","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1177/20965311231168423","year":2023,"title":"ECNU Review of Education6(3), 355–366 (Aug 2023)","work_id":"de907ac7-f6d3-44a4-8a0b-d05ecc023231","ref_index":6,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1080/02601370.2024.2310448","year":2024,"title":"Milana, M., Brandi, U., Hodge, S., & Hoggan-Kloubert, T. (2024). Artificial intelligence (AI), conversational agents, and generative AI: implications for adult education practice and research. Interna","work_id":"5efaa1a3-1e60-4b04-b258-958fa2dd9bfd","ref_index":7,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":144,"snapshot_sha256":"4afb8c87bbced638327bb16338190caba31d272a4d4054a236b78ac2556bb230","internal_anchors":1},"formal_canon":{"evidence_count":2,"snapshot_sha256":"09151990cdfb154bd98735f9ea36252711a9b5c14fe62e5e85b22b5b6dfb8eaa"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}