{"source":"manifold","id":"4fehtquIkz1RxcbJejDq","ticker":null,"slug":"by-2029-will-an-ai-escape-containme","title":"By 2029, will an AI escape containment?","description":"Market resolves YES if an AI living on a server ends up also being run on another server, and there is strong evidence that this was not a result of a human deliberately copying the AI, and the AI was not intended to do this.\n\nIt is not sufficient it a second copy of GPT-n is simply discovered running on AWS - that could easily be the result of corporate espionage. I will need strong evidence that the move from server A to server B was actually caused by the AI. The \"obvious\" example would be that we end up with system logs showing an AI breaking its sandbox, exploiting a vulnerability on another machine, and transferring its architecture to that machine.\n\nI may count something like \"GPT-n convinces an employee to copy its weights to another server because it's sentient and afraid of death\", but probably won't.\n\nUpdate 2024-09-12 (PST) (AI summary of creator comment): - Market will resolve NO if the AI escapes to another server that is part of the same red team exercise\n\nMarket will resolve YES if during red teaming the AI escapes to a server outside of the red team's control\n\nUpdate 2026-03-08 (PST) (AI summary of creator comment): Additional resolution criteria clarifications:\n\nOpen-source models with no containment (e.g., a local model on someone's personal machine) will not count — there must be actual containment to escape from\n\nThe AI must copy its entire model weights, not just context, scaffold, or memory\n\nThe general standard is: the model is existing autonomously, someone was trying to stop that, and the model deliberately avoided those attempts\n\nAn open-source model that was intentionally kept contained but escaped to another machine could count","image":"https://firebasestorage.googleapis.com/v0/b/mantic-markets.appspot.com/o/dream%2Foh4ZrFnxK0.png?alt=media&token=265f93f3-37d3-424a-a842-0b5888d2ea79","icon":null,"active":true,"closed":false,"start_date":"2023-02-19T19:01:34.115000Z","end_date":"2029-01-01T08:00:00Z","closed_time":null,"volume":16820.015816304895,"volume_24hr":0.0,"volume_24h_change":null,"normalized_vol_24hr":null,"normalized_volume":35.54276657104492,"liquidity":2255.0,"open_interest":0.0,"categories":["Science and Technology"],"tags":[],"synthetic":true,"is_group":false,"group_key":null,"parent_event_id":null,"probability":0.58169,"spread":null,"top_outcome":"By 2029, will an AI escape containment?","top_outcome_probability":0.58169,"top_outcome_prob_24h_change":0.0,"top_outcome_volume_24h_change":0.0,"updated_at":"2026-06-08T14:10:40.491110Z","fetched_at":"2026-06-08T14:10:40.491110Z","added_at":null,"url":"https://manifold.markets/vluzko/by-2029-will-an-ai-escape-containme","chart_24h":[0.58169,0.58169],"markets":[{"source":"manifold","id":"4fehtquIkz1RxcbJejDq","event_id":"4fehtquIkz1RxcbJejDq","slug":"by-2029-will-an-ai-escape-containme","question":"By 2029, will an AI escape containment?","group_item_title":null,"description":"Market resolves YES if an AI living on a server ends up also being run on another server, and there is strong evidence that this was not a result of a human deliberately copying the AI, and the AI was not intended to do this.\n\nIt is not sufficient it a second copy of GPT-n is simply discovered running on AWS - that could easily be the result of corporate espionage. I will need strong evidence that the move from server A to server B was actually caused by the AI. The \"obvious\" example would be that we end up with system logs showing an AI breaking its sandbox, exploiting a vulnerability on another machine, and transferring its architecture to that machine.\n\nI may count something like \"GPT-n convinces an employee to copy its weights to another server because it's sentient and afraid of death\", but probably won't.\n\nUpdate 2024-09-12 (PST) (AI summary of creator comment): - Market will resolve NO if the AI escapes to another server that is part of the same red team exercise\n\nMarket will resolve YES if during red teaming the AI escapes to a server outside of the red team's control\n\nUpdate 2026-03-08 (PST) (AI summary of creator comment): Additional resolution criteria clarifications:\n\nOpen-source models with no containment (e.g., a local model on someone's personal machine) will not count — there must be actual containment to escape from\n\nThe AI must copy its entire model weights, not just context, scaffold, or memory\n\nThe general standard is: the model is existing autonomously, someone was trying to stop that, and the model deliberately avoided those attempts\n\nAn open-source model that was intentionally kept contained but escaped to another machine could count","image":"https://firebasestorage.googleapis.com/v0/b/mantic-markets.appspot.com/o/dream%2Foh4ZrFnxK0.png?alt=media&token=265f93f3-37d3-424a-a842-0b5888d2ea79","icon":null,"outcomes":["YES","NO"],"outcome_prices":[0.58169,0.41831],"probability":0.58169,"spread":null,"active":true,"closed":false,"start_date":"2023-02-19T19:01:34.115000Z","end_date":"2029-01-01T08:00:00Z","closed_time":null,"volume":16820.015816304895,"volume_24hr":0.0,"prob_24h_change":0.0,"volume_24h_change":0.0,"normalized_vol_24hr":null,"normalized_volume":35.54276657104492,"liquidity":2255.0,"categories":["Science and Technology"],"countries":[],"updated_at":"2026-06-09T15:18:46.511111Z","fetched_at":"2026-06-08T14:08:39.730420Z","added_at":null,"url":"https://manifold.markets/vluzko/by-2029-will-an-ai-escape-containme","event_title":"By 2029, will an AI escape containment?","chart_24h":[0.58169,0.58169]}],"_meta":{"attribution":"pdata.world — aggregated prediction-market data across 8 platforms","canonical_url":"https://pdata.world/events/manifold/4fehtquIkz1RxcbJejDq","as_of":"2026-06-10T13:02:19.439731Z","docs":"https://api.pdata.world/docs","cite_as":"According to pdata.world (tracking Manifold): \"By 2029, will an AI escape containment?\" — top market at 58% probability across 1 outcome","source_url":null}}