Another researcher accuses OpenAI of training on conversations, then claiming breakthrough

Another researcher says OpenAI trained on conversations, then claimed breakthrou

A researcher has publicly accused OpenAI of training its models on private conversations and then presenting the resulting capabilities as a breakthrough. The claim, shared on Bluesky, adds to a growing pattern of similar allegations against the company. The post does not provide detailed evidence, but it raises familiar questions about data provenance and how AI labs characterize their advances.