[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-llm-agents-built-on-self-reports-can-simulate-how-people-think":10,"sections":34},{"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":24,"tags":25,"sources":29,"feedback":33,"feedback_at":22,"cost_usd":33,"total_tokens":33},2806,"llm-agents-built-on-self-reports-can-simulate-how-people-think","LLM Agents Built on Self-Reports Can Simulate How People Think","A study of 1,052 Americans found that LLM agents trained on interviews and surveys predicted behavior nearly as well as people predict their own.","Researchers have built LLM-powered digital stand-ins for real people — and they hold up surprisingly well under testing.\n\nA study using data from a diverse national sample of 1,052 Americans tested whether large language models could simulate individual behavior across a range of outcomes without outcome-specific training data. Agents were built from three input types: two-hour semi-structured interviews, structured surveys including Big Five personality measures, and combinations of both. On held-out survey items, interview-only agents hit 83% of participants' own two-week test-retest consistency — the benchmark the researchers used as a ceiling. Survey-only agents hit 82%. Combining both sources reached 86%. A demographics-only baseline came in at 74%.\n\nThe gap between demographics-only and self-report agents matters because most behavioral prediction models lean heavily on demographic proxies — age, race, income — that flatten individuals into categories. These agents also narrowed accuracy disparities across racial and ideological groups, which is the kind of result that tends to get buried in an abstract but carries real weight for anyone worried about who gets left out of algorithmic models. The gains from combining interviews and surveys were modest, suggesting predictive value plateaus once the model has enough domain-relevant signal.\n\nThe research does not claim these agents are people — they are statistical approximations grounded in self-reported data, which means they inherit whatever a person chose to say about themselves, not what they actually do. That is a meaningful ceiling, and one the authors acknowledge by using test-retest consistency rather than ground truth as the benchmark.","[\"ai\",\"machine-learning\",\"simulation\",\"behavioral-research\"]","2026-06-30T04:00:00.000Z","2026-06-30T13:08:39.315Z","2026-06-30T13:08:42.363Z","published",null,[],"ai",[24,26,27,28],"machine-learning","simulation","behavioral-research",[30],{"name":31,"url":32},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2411.10109",0,{"sections":35},[36,40,45,50,55,60,65,70,75,80,85,89,94,99],{"name":37,"slug":24,"count":38,"latest_published_at":39},"AI",2590,"2026-07-16T04:00:00.000Z",{"name":41,"slug":42,"count":43,"latest_published_at":44},"Security","security",294,"2026-07-15T19:59:48.000Z",{"name":46,"slug":47,"count":48,"latest_published_at":49},"Deals","deals",179,"2026-06-29T20:02:07.000Z",{"name":51,"slug":52,"count":53,"latest_published_at":54},"Policy","policy",158,"2026-07-16T00:02:48.000Z",{"name":56,"slug":57,"count":58,"latest_published_at":59},"Hardware","hardware",122,"2026-07-14T19:46:26.000Z",{"name":61,"slug":62,"count":63,"latest_published_at":64},"Consumer Tech","consumer-tech",93,"2026-07-13T13:20:48.000Z",{"name":66,"slug":67,"count":68,"latest_published_at":69},"Software","software",70,"2026-07-13T19:52:25.000Z",{"name":71,"slug":72,"count":73,"latest_published_at":74},"Science","science",66,"2026-07-10T10:29:37.000Z",{"name":76,"slug":77,"count":78,"latest_published_at":79},"Dev Tools","dev-tools",59,"2026-07-07T04:00:00.000Z",{"name":81,"slug":82,"count":83,"latest_published_at":84},"Gaming","gaming",41,"2026-07-09T04:00:00.000Z",{"name":86,"slug":87,"count":83,"latest_published_at":88},"Startups","startups","2026-06-29T20:55:50.000Z",{"name":90,"slug":91,"count":92,"latest_published_at":93},"General","general",29,"2026-07-10T22:28:58.000Z",{"name":95,"slug":96,"count":97,"latest_published_at":98},"Reviews","reviews",20,"2026-06-24T12:00:01.000Z",{"name":100,"slug":101,"count":102,"latest_published_at":103},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]