Fired OpenAI researchers ask the company to preserve visibility into AI reasoning

The Wall Street Journal reported in an exclusive on October 7, 2026 that a group of researchers who were fired by OpenAI are asking the company to preserve visibility into its AI reasoning.

The request centers on whether the intermediate steps a model generates before its final answer, commonly called its chain of thought, remain readable and reviewable from the outside. For these researchers, that is not a product nicety but a precondition for independently assessing model behaviour and catching risks early.

The appeal lands as the industry debates just how opaque frontier reasoning has become. As leading models make it harder to observe their thinking in full, safety researchers and external auditors worry that once that reasoning is hidden from monitors, it becomes much harder to tell whether a system might deceive, evade oversight or otherwise behave dangerously.

The demand speaks to a tension between model transparency and commercial secrecy. Companies often restrict access to reasoning traces in the name of safety, abuse prevention or intellectual property, while safety researchers argue that closing it off can leave external oversight with nothing to inspect.

Notably, the researchers are making the case after leaving the company, which frees them from internal channels and turns a debate that once played out inside OpenAI into a public one, putting the transparency question squarely in the wider safety conversation.

It is not yet clear how OpenAI will respond. A commitment to keep reasoning visible could set a transparency benchmark for the industry, while maintaining the status quo risks deepening doubts about the company's safety pledges, making OpenAI's answer and any regulatory follow-up the things to watch.