Another researcher says OpenAI trained on conversations, then claimed breakthrough. The claim suggests the company used private or public chat data to develop its models before presenting results as novel advancements. This follows ongoing scrutiny over OpenAI's training data sources and transparency.

Critics argue that using conversational data without clear disclosure undermines claims of originality. The researcher's statement adds to growing concerns about ethical AI development and data provenance. No direct evidence is provided in the summary, but the allegation aligns with prior debates about model training practices.

The discussion appeared on Hacker News, drawing 76 points and 30 comments. The original post was shared via Bluesky, linking to a researcher's profile. The core issue remains whether OpenAI adequately disclosed its data usage.

Such claims could impact trust in AI benchmarks and research integrity. The debate highlights tensions between innovation and accountability in AI development.