![\"\"](\"https:\/\/www.newsbeep.com\/uk\/wp-content\/uploads\/2026\/06\/SEI_302080413.jpg\")
<\/p>\n<\/p>\n
Having one chatbot train another could be a recipe for disaster<\/p>\n
fotograzia\/Getty Images<\/p>\n<\/p>\n
People who are paid to train new AI models by supplying them with high-quality conversation and tests are cheating and using chatbots like ChatGPT to do the job instead, multiple whistleblowers have told New Scientist. The seemingly widespread practice risks undermining the future of AI, as it could lead to the \u201ccollapse\u201d of more advanced models.<\/p>\n
Most AI models operating today were trained on text and data scraped from the internet<\/a>. But as models have scaled up, requiring yet more training data, AI firms have begun using workers who carry out conversations and tests with AI, in the hope that the resulting high-quality data can improve the power and usefulness of future large language models (LLMs).<\/p>\n These workers are normally employed by third parties, rather than AI companies directly, and are often working without full-time contracts and for low pay. That can incentivise them to take shortcuts like using chatbots to complete tasks faster, according to a worker called Alice*, despite this being against company policies.<\/p>\n \u201cIt\u2019s very widespread; every company I\u2019ve worked for has had explicit guidelines around it and they clearly do try to catch people out, so I think they do care. But I don\u2019t think they can stop it,\u201d says Alice.<\/p>\n Alice says she\u00a0feels \u201cnot in the slightest\u201d guilty about using ChatGPT to complete training tasks, saying it is easy to get away with as long as you instruct chatbots to avoid the usual telltale signs of AI output, like a preponderance of em-dashes<\/a>. \u201cIt\u2019s only the sloppiest of users that get caught,\u201d she says. \u201cAnyone with a modicum of awareness around AI hallmarks can tell their output not to use them, and at that point what are you going to do?\u201d<\/p>\n \u201cIf these companies want quality data, then they should offer quality contracts,\u201d says Alice. \u201cInstead they\u2019re low-balling struggling people, employing them for the barest possible amount of time and tossing them aside as projects are finished with no warning.\u201d<\/p>\n Another worker, Bob*, worked for a training platform called Outlier. Initially, he was tasked with AI training, which he says he illicitly used AI for, and was then promoted to a leadership role where part of his job was to catch others doing the same thing.<\/p>\n \u201cManagement vacillated between light tolerance to outright banning,\u201d says Bob. Workers at Outlier would be tracked with a tool called Hubstaff which takes screenshots of their desktop at random intervals to ensure they are really doing tasks as ordered. Bob would look for evidence of AI models in those screenshots.<\/p>\n \u201cPeople would have it [AI models like ChatGPT] open in other tabs, or minimised, so obviously we could see it in the task bar,\u201d says Bob. \u201cEven stuff like folders on their desktop with names gave it [AI use] away.\u201d<\/p>\n Outlier, which is owned by Scale AI, did not respond to a request for comment. Scale AI claims on its website<\/a> to carry out work for technology giants like Meta and Cisco, neither of which responded to New Scientist\u2018s request for comment. Bob says he had personally worked on projects for Google, which also did not respond to a request for comment.<\/p>\n Another worker, Carol*, who has worked on several platforms, says that her use of AI began by checking her work for anything that went against the lengthy guidelines for a task, because any contravention could mean expulsion from the project and a loss of earnings.<\/p>\n \u201cI was terrified of not having an income source, and then after that, it just became easier to run everything through LLMs,\u201d says Carol. \u201cFor a lot of the projects that I do now, it\u2019s creating scenarios, so I will use one LLM to help me create the scenario and then I\u2019ll use a different LLM to help me create the files that go along with the scenario. I do feel guilty but like I said, in the beginning it was more about trying to make sure I wasn\u2019t making any errors.\u201d<\/p>\n \u201cI do worry that I\u2019m actually making it [AI] worse. I thought using the models to train themselves negates some of the value,\u201d says Carol.<\/p>\n Mark Lee<\/a> at the University of Birmingham, UK, says research has shown that AI models \u201ccollapse\u201d if they are recursively trained on AI-generated content<\/a>. When this happens, the abilities of the model drop dramatically and they become less useful. The process is sometimes known as AI cannibalism or AI inbreeding.<\/p>\n \u201cThat\u2019s the kind of worst-case scenario. And that\u2019s probably not what\u2019s happening in the real world,\u201d says Lee. \u201cThere\u2019s still a few humans. And if you have like 10 per cent human data, it mitigates it, it avoids model collapse.\u201d<\/p>\n But Lee says that the kind of cheating these workers are doing isn\u2019t without repercussions, and will hit performance. \u201cRather than it being catastrophic, you\u2019ll see that the AI isn\u2019t as good at doing human-like tasks. It\u2019s an issue, because I think the models aren\u2019t as good as they could be.\u201d<\/p>\n *Names have been changed to protect identities<\/p>\n Topics:<\/p>\n","protected":false},"excerpt":{"rendered":"Having one chatbot train another could be a recipe for disaster fotograzia\/Getty Images People who are paid to…\n","protected":false},"author":2,"featured_media":662293,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[554,733,4308,86,56,54,55],"class_list":["post-662292","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-artificial-intelligence","tag-artificialintelligence","tag-technology","tag-uk","tag-united-kingdom","tag-unitedkingdom"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/662292","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/comments?post=662292"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/662292\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media\/662293"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media?parent=662292"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/categories?post=662292"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/tags?post=662292"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}