{"id":1188305,"date":"2026-10-01T14:33:52","date_gmt":"2026-10-01T21:33:52","guid":{"rendered":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/publication\/use-and-effects-of-llms-in-peer-review-a-randomized-experiment-and-survey-at-icml-2026\/"},"modified":"2026-10-07T13:28:51","modified_gmt":"2026-10-07T20:28:51","slug":"use-and-effects-of-llms-in-peer-review-a-randomized-experiment-and-survey-at-icml-2026","status":"publish","type":"msr-research-item","link":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/publication\/use-and-effects-of-llms-in-peer-review-a-randomized-experiment-and-survey-at-icml-2026\/","title":{"rendered":"Use and Effects of LLMs in Peer Review: A Randomized Experiment and Survey at ICML 2026"},"content":{"rendered":"\n\n\n<p class=\"wp-block-paragraph\">LLMs are rapidly reshaping peer review, making it important to understand how reviewers use them in practice and how different LLM-use policies affect review outcomes. We investigate these questions through a randomized experiment and an anonymous post-survey at ICML 2026, a major machine learning conference involving over 24,000 papers and 17,000 reviewers. Reviewers were assigned to either a conservative policy prohibiting all LLM use or a permissive policy allowing limited assistance, with randomization among a subset of main-track papers and reviewers. Policy assignment had near-zero effects on final paper decisions, paper scores, and reviewer confidence, although reviews under the permissive policy were 5.5-7% longer. Post-survey responses (N=1,486) revealed diverse attitudes toward LLMs and substantial noncompliance: 22.5% of conservative-policy reviewers reported using an LLM despite the prohibition, and 36.5% of permissive-policy reviewers reported at least one explicitly disallowed use. We discuss implications for future peer-review policy and tool design.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>LLMs are rapidly reshaping peer review, making it important to understand how reviewers use them in practice and how different LLM-use policies affect review outcomes. We investigate these questions through a randomized experiment and an anonymous post-survey at ICML 2026, a major machine learning conference involving over 24,000 papers and 17,000 reviewers. Reviewers were assigned [&hellip;]<\/p>\n","protected":false},"featured_media":0,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","msr-author-ordering":[{"type":"user_nicename","value":"Sunnie S. Y. Kim","user_id":"44180"},{"type":"user_nicename","value":"Wesley Deng","user_id":"44164"},{"type":"user_nicename","value":"Jenn Wortman Vaughan","user_id":"32235"},{"type":"text","value":"Buxin Su","user_id":0},{"type":"text","value":"Wei-Jie Su","user_id":0},{"type":"user_nicename","value":"Alekh Agarwal","user_id":"30928"},{"type":"text","value":"Sharon Li","user_id":0},{"type":"text","value":"Martin Jaggi","user_id":0},{"type":"user_nicename","value":"Dan Goldstein","user_id":"31618"},{"type":"text","value":"Nihar B. Shah","user_id":0},{"type":"user_nicename","value":"Miro Dud\u00edk","user_id":"32867"}],"msr_publishername":"","msr_publisher_other":"","msr_booktitle":"","msr_chapter":"","msr_edition":"","msr_editors":"","msr_how_published":"arXiv","msr_isbn":"","msr_issue":"","msr_journal":"","msr_number":"","msr_organization":"","msr_pages_string":"","msr_page_range_start":"","msr_page_range_end":"","msr_series":"","msr_volume":"","msr_copyright":"","msr_conference_name":"","msr_doi":"","msr_arxiv_id":"2609.19420","msr_mag_id":"","msr_other_authors":"","msr_other_contributors":"","msr_speaker":"","msr_award":"","msr_affiliation":"","msr_institution":"","msr_host":"","msr_version":"","msr_duration":"","msr_release_tracker_id":"","msr_highlight_type":"","msr_date_display_format":"","msr_main_download_label":"","msr_external_link_label":"","msr_doi_label":"","msr_published_date":"2026-09-16","msr_startdate":"","msr_presentation_date":"","msr_highlight_text":"","msr_notes":"","msr_longbiography":"","msr_publicationurl":"","msr_external_url":"","msr_secondary_video_url":"","msr_conference_url":"","msr_journal_url":"","msr_year":2026,"msr_month":9,"msr_day":16,"msr_microsoftintellectualproperty":false,"msr_pub_id":"8f8bc525ebccecd1f8af7c67e17d562d577bff6d","msr_publication_uploader":[{"type":"url","title":"https:\/\/arxiv.org\/abs\/2609.19420","label_id":243109,"id":false,"viewUrl":false}],"msr_related_uploader":[],"msr_original_fields_of_study":[],"msr_s2_paper_id":"8f8bc525ebccecd1f8af7c67e17d562d577bff6d","msr_s2_pdf_url":"","msr_citation_count_updated":"","msr_citation_count":0,"msr_influential_citations":0,"msr_reference_count":82,"msr_s2_open_access":false,"msr_s2_author_ids":[],"msr_pub_ids":[{"provider":"s2","id":"8f8bc525ebccecd1f8af7c67e17d562d577bff6d"},{"provider":"arxiv","id":"2609.19420"},{"provider":"corpusid","id":"292059243"}],"msr_hide_image_in_river":0,"footnotes":""},"msr-research-highlight":[],"research-area":[13554],"msr-publication-type":[270373],"msr-publisher":[],"msr-publication-cta":[],"msr-focus-area":[],"msr-locale":[268875],"msr-post-option":[],"msr-field-of-study":[246691,255217],"msr-conference":[],"msr-journal":[],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-1188305","msr-research-item","type-msr-research-item","status-publish","hentry","msr-research-area-human-computer-interaction","msr-locale-en_us","msr-field-of-study-computer-science","msr-field-of-study-human-computer-interaction-4"],"msr_publishername":"","msr_edition":"","msr_affiliation":"","msr_published_date":"2026-09-16","msr_host":"","msr_duration":"","msr_version":"","msr_speaker":"","msr_other_contributors":"","msr_booktitle":"","msr_pages_string":"","msr_chapter":"","msr_isbn":"","msr_journal":"","msr_volume":"","msr_number":"","msr_editors":"","msr_series":"","msr_issue":"","msr_organization":"","msr_how_published":"arXiv","msr_notes":"","msr_highlight_text":"","msr_release_tracker_id":"","msr_original_fields_of_study":"","msr_download_urls":"","msr_external_url":"","msr_secondary_video_url":"","msr_longbiography":"","msr_microsoftintellectualproperty":0,"msr_main_download":"","msr_publicationurl":"","msr_doi":"","msr_publication_uploader":[{"type":"url","title":"https:\/\/arxiv.org\/abs\/2609.19420","label_id":243109,"id":false,"viewUrl":false}],"msr_related_uploader":[],"msr_citation_count":0,"msr_citation_count_updated":"","msr_s2_paper_id":"8f8bc525ebccecd1f8af7c67e17d562d577bff6d","msr_influential_citations":0,"msr_reference_count":82,"msr_arxiv_id":"2609.19420","msr_s2_author_ids":[],"msr_s2_open_access":false,"msr_s2_pdf_url":null,"msr_attachments":[],"msr-author-ordering":[{"type":"user_nicename","value":"Sunnie S. Y. Kim","user_id":44180,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Sunnie S. Y. Kim"},{"type":"user_nicename","value":"Wesley Deng","user_id":44164,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Wesley Deng"},{"type":"user_nicename","value":"Jenn Wortman Vaughan","user_id":32235,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Jenn Wortman Vaughan"},{"type":"text","value":"Buxin Su","user_id":0,"rest_url":false},{"type":"text","value":"Wei-Jie Su","user_id":0,"rest_url":false},{"type":"user_nicename","value":"Alekh Agarwal","user_id":30928,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Alekh Agarwal"},{"type":"text","value":"Sharon Li","user_id":0,"rest_url":false},{"type":"text","value":"Martin Jaggi","user_id":0,"rest_url":false},{"type":"user_nicename","value":"Dan Goldstein","user_id":31618,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Dan Goldstein"},{"type":"text","value":"Nihar B. Shah","user_id":0,"rest_url":false},{"type":"user_nicename","value":"Miro Dud\u00edk","user_id":32867,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Miro Dud\u00edk"}],"msr_impact_theme":[],"msr_research_lab":[199563,199571],"msr_event":[],"msr_group":[],"msr_project":[],"publication":[],"video":[],"msr-tool":[],"msr_publication_type":"misc","related_content":[],"_links":{"self":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/1188305","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item"}],"about":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-research-item"}],"version-history":[{"count":4,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/1188305\/revisions"}],"predecessor-version":[{"id":1189011,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/1188305\/revisions\/1189011"}],"wp:attachment":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/media?parent=1188305"}],"wp:term":[{"taxonomy":"msr-research-highlight","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-highlight?post=1188305"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=1188305"},{"taxonomy":"msr-publication-type","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-type?post=1188305"},{"taxonomy":"msr-publisher","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publisher?post=1188305"},{"taxonomy":"msr-publication-cta","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-cta?post=1188305"},{"taxonomy":"msr-focus-area","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-focus-area?post=1188305"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=1188305"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=1188305"},{"taxonomy":"msr-field-of-study","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-field-of-study?post=1188305"},{"taxonomy":"msr-conference","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-conference?post=1188305"},{"taxonomy":"msr-journal","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-journal?post=1188305"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=1188305"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=1188305"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}