{"slug": "openai-wants-you-to-use-ai-but-not-to-train-its-ai", "title": "OpenAI wants you to use AI — but not to train its AI", "summary": "OpenAI fired an undisclosed number of contractors who were hired to read ChatGPT users' prompts and provide human feedback after discovering many were using AI tools to do the work, according to a 404 Media report. The contractors' working conditions explicitly banned AI use, including Grammarly and AI translation, warning that \"Reviewers may not use AI either.\" OpenAI declined to comment, and the company is seeking to avoid \"model collapse,\" in which models trained on AI-generated text perform worse than their predecessors.", "body_md": "Here’s an interesting concept: an AI company that fires people for using AI. It sounds like a strange way to run a company but there’s a real reason behind it.\n\nOpenAI has been hiring contractors who have been tasked with reading ChatGPT users’ prompts to help improve responses by providing some real human input.\n\nUnfortunately for OpenAI, it found many of them were using AI to train the AI, so not providing a human touch at all, [according to a report in 404 Media](https://www.404media.co/people-training-openais-ai-fired-for-using-ai-to-train-the-ai/). The company has subsequently fired many of the contractors, although 404 Media didn’t reveal the number.\n\nNone of them can say they weren’t warned, however. The terms for the contractors are set out in their working conditions. “Do not use AI detection tools, or AI yourself. Do not use GPTZero or any other AI detection tool. They are not reliable. Reviewers may not use AI either, including Grammarly and AI translation, to review, write feedback, or write comments.”\n\nDespite this stark warning, many of the contractors turned to AI to assist in the work. OpenAI is very keen to avoid “model collapse,” a phenomenon in which AI models trained on text written by previous generations of AI models perform worse than their predecessors. It’s a form of digital inbreeding that [industry observers have previously warned about,](https://www.infoworld.com/article/3988468/what-comes-after-stack-overflow.html#:~:text=This%20creates%20the%20risk%20of%20model%20collapse%2C%20where%20AI%20increasingly%20recycles%20its%20outputs%20rather%20than%20ingesting%20new%2C%20human%2Dvalidated%20insights.) and [which could have a deleterious impact on business](https://www.cio.com/article/4151360/the-hidden-inflation-of-ai-why-model-collapse-is-a-business-risk.html).\n\nOne contractor told 404 Media that it was a common practise. “It’s pretty much the one thing that will get you kicked off ASAP. In a group of thousands, there are tons that have been caught.”\n\nIt will be hard to ascertain just how widespread the practise is as OpenAI declined to comment on the situation.", "url": "https://wpnews.pro/news/openai-wants-you-to-use-ai-but-not-to-train-its-ai", "canonical_source": "https://www.computerworld.com/article/4226841/openai-wants-you-to-use-ai-but-not-to-train-its-ai.html", "published_at": "2026-09-25 17:27:15+00:00", "updated_at": "2026-09-25 17:29:55.856291+00:00", "lang": "en", "topics": ["artificial-intelligence", "large-language-models", "ai-safety"], "entities": ["OpenAI", "ChatGPT", "404 Media", "Grammarly", "GPTZero"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/openai-wants-you-to-use-ai-but-not-to-train-its-ai", "markdown": "https://wpnews.pro/news/openai-wants-you-to-use-ai-but-not-to-train-its-ai.md", "text": "https://wpnews.pro/news/openai-wants-you-to-use-ai-but-not-to-train-its-ai.txt", "jsonld": "https://wpnews.pro/news/openai-wants-you-to-use-ai-but-not-to-train-its-ai.jsonld"}}