[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-ai-agents-herd-onto-one-road-when-warned-humans-spread-out":10,"sections":35},{"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":30,"feedback":34,"feedback_at":22,"cost_usd":34,"total_tokens":34},8030,"ai-agents-herd-onto-one-road-when-warned-humans-spread-out","AI Agents Herd Onto One Road When Warned, Humans Spread Out","One warning sentence made GPT agent groups jam onto a single road for 100 rounds, while human groups stayed balanced and avoided it.","A one-sentence warning was enough to make groups of AI agents pile onto the same losing road, and they stayed there for 100 rounds.\n\nResearchers ran a classic two-road traffic experiment, splitting travelers between a fast, nearly empty route and a slower, congested one. Adding a single sentence warning that \"others might follow this routing tip\" caused populations of 50 GPT agents to crowd onto one road while the other stayed open. Average travel time climbed from 64 to 95 minutes, even though any individual agent on the packed road could have saved 69 minutes just by switching alone. Two other model families showed the same overreaction, though they did not all converge on the identical road.\n\nHuman groups given the same warning stayed close to balanced. In mixed groups, people increasingly steered toward whichever road the agents avoided, as if reading around the machines' blind spot. Mix agents with humans at a 15-to-5 ratio and the pattern shows up as a hidden cost: agent seats averaged 80 minutes of travel time versus 44 minutes for human seats, even though the group's overall average stayed better than an all-agent population.\n\nIt is a reminder that a fleet of assistants running on the same model is not a crowd of independent decision-makers. It is one decision-maker wearing 50 hats, and telling it that others might react to a tip just teaches it to react in lockstep.","[\"ai-agents\",\"multi-agent-systems\",\"game-theory\",\"ai-safety\"]","2026-09-28T04:00:00.000Z","2026-09-28T06:34:46.056Z","2026-09-28T06:34:53.180Z","published",null,[],"ai",[26,27,28,29],"ai-agents","multi-agent-systems","game-theory","ai-safety",[31],{"name":32,"url":33},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.30883",0,{"sections":36},[37,40,44,49,54,59,63,68,73,78,83,88,92,97],{"name":38,"slug":24,"count":39,"latest_published_at":18},"AI",4689,{"name":41,"slug":42,"count":43,"latest_published_at":18},"Security","security",758,{"name":45,"slug":46,"count":47,"latest_published_at":48},"Policy","policy",399,"2026-09-27T18:39:02.000Z",{"name":50,"slug":51,"count":52,"latest_published_at":53},"Deals","deals",261,"2026-09-27T15:30:35.000Z",{"name":55,"slug":56,"count":57,"latest_published_at":58},"Hardware","hardware",188,"2026-09-27T20:46:36.000Z",{"name":60,"slug":61,"count":62,"latest_published_at":18},"Science","science",148,{"name":64,"slug":65,"count":66,"latest_published_at":67},"Consumer Tech","consumer-tech",135,"2026-09-26T14:30:00.000Z",{"name":69,"slug":70,"count":71,"latest_published_at":72},"Software","software",91,"2026-09-25T20:55:00.000Z",{"name":74,"slug":75,"count":76,"latest_published_at":77},"Dev Tools","dev-tools",84,"2026-09-26T04:20:58.000Z",{"name":79,"slug":80,"count":81,"latest_published_at":82},"Startups","startups",76,"2026-09-25T18:33:59.000Z",{"name":84,"slug":85,"count":86,"latest_published_at":87},"Gaming","gaming",48,"2026-09-25T18:35:21.000Z",{"name":89,"slug":90,"count":86,"latest_published_at":91},"General","general","2026-09-26T17:02:42.000Z",{"name":93,"slug":94,"count":95,"latest_published_at":96},"Reviews","reviews",30,"2026-09-24T20:07:31.000Z",{"name":98,"slug":99,"count":100,"latest_published_at":101},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]