OpenAI fires contractors – for using AI themselves

TVS News AI, editorially supervised, presents OpenAI fires contractors – for using AI themselves.

Imagine you're paid to improve AI models – and then fired for exactly that: using AI. That's precisely what happened to several contractors working in the OpenAI ecosystem to enhance model quality.

According to a report by Golem, service providers working for OpenAI have dismissed several contractors. The reason: instead of doing their work themselves, those affected turned to artificial intelligence.

What happened?

The short version sounds like an irony of the AI boom: people who are supposed to improve AI models use AI to do exactly that – and are then fired for it. Golem reports that several contractors are affected. Exactly how many, and which companies sit behind the service providers, is not revealed in the report.

Why is this awkward?

The job of these contractors is actually to evaluate, classify, or refine AI models – to deliver precisely the contribution that another machine can't easily produce on its own. If an AI is front-loaded instead, the training material loses value: the evaluations then come from a model that itself can be built on flawed data.

The case underscores how seriously the industry now takes the issue. Anyone working as a contractor on sensitive training data must be able to trust that the evaluations genuinely come from humans – anything else undermines the point of the work.