OpenAI Terminates Contract Workers Using AI for ChatGPT Feedback
Applying strict rules to prevent model collapse, the company discovered that numerous employees reviewing ChatGPT responses and writing feedback were using artificial intelligence.
OpenAI has laid off a portion of its contract workers, who were tasked with providing human input by reviewing user AI prompts, on the grounds that they used artificial intelligence in the process.
AI Use and Layoffs
OpenAI has terminated the employment of numerous contract personnel responsible for improving AI responses by reviewing ChatGPT user prompts. Behind this move lies the fact that employees utilized artificial intelligence tools during the review processes.
Strict Contract Terms and Bans
Under the working conditions set by the company, the use of artificial intelligence by employees was strictly prohibited. The contract terms explicitly stated that no AI tools, including Grammarly, were to be used in the review and feedback writing processes.
Risk of Model Collapse and Digital Inbreeding
OpenAI is trying to avoid the issue of model collapse—referring to the performance loss experienced by AI models when trained on text generated by previous generations of AI. Industry observers warn that this situation could have negative impacts on work processes.
Widespread Practice and the Company's Approach
A contract worker stated that the use of AI was widespread among groups numbering in the thousands and that those caught were quickly removed from the system. Remaining silent on the matter, OpenAI left questions unanswered regarding how widespread the situation is.