logo naukrinama

OpenAI's Strict Measures Against Contractors Misusing AI Tools

OpenAI has recently taken stringent measures against contract workers who misused AI tools while training its models. A report revealed that some workers, instead of providing human-like feedback, relied on AI for their tasks, leading to immediate terminations. The company emphasizes the critical role of human judgment in AI training to prevent model collapse. Despite clear guidelines against using AI detection tools, some workers still attempted to circumvent these rules. This article explores the implications of these actions and the experiences of those affected.
By Ambika Singh Sep 25, 2026, 13:35 IST

OpenAI's Approach to Humanizing AI Models

New Delhi: To enhance the human-like qualities of its AI models, OpenAI employs thousands of contract workers. However, a startling revelation from a media report has surfaced. Several workers, tasked with teaching AI to mimic human interaction, were found using AI tools to complete their assignments. Following this discovery, the company promptly terminated the employment of those involved.


Understanding the Project's Functionality

According to the report, these contract workers analyze conversations and prompts from actual ChatGPT users to assess the accuracy and balance of the chatbot's responses. Their primary responsibility is to ensure that the AI does not overly agree with users, avoiding an excessively flattering demeanor. For this role, they earn approximately $50 per hour.


The Challenge of Monitoring Thousands of Workers

Despite the involvement of thousands of contract workers in this project, many opted to simplify their tasks by utilizing AI tools, which directly contradicts company policy. The report indicates that workers caught in such violations are swiftly dismissed, a practice that has become commonplace.


The Risks of Training AI with AI-Generated Content

OpenAI is enforcing strict measures because training AI models with AI-generated content poses a risk of 'model collapse.' This phenomenon can lead to a decline in the model's accuracy and diversity over time. Hence, human insight and judgment are deemed crucial in this process, something that machines cannot replicate.


A Worker’s Heartbreaking Experience

The report mentions a worker who received a termination letter questioning the 'authenticity' of their work. This individual expressed that they were not a bad person or employee; they simply sought assistance from AI. They also noted that they found no joy or satisfaction in their work.


Strict Guidelines Yet Some Noncompliance

The company has issued clear instructions prohibiting workers from using any AI detection or writing tools, as these are not entirely reliable. Reviewers have also been advised against using tools like Grammarly or translation services for feedback writing. Instead, they are instructed to pay attention to repeated phrases and unique punctuation styles as indicators.


FROM AROUND THE WEB