The New York Times escalated its copyright battle with OpenAI on Wednesday, filing a motion for sanctions that accuses the company of hiding evidence. Publishers allege OpenAI concealed tools and datasets capable of identifying copyrighted journalism within ChatGPT outputs. The motion claims the AI firm deleted or failed to preserve billions of relevant logs.
This moves beyond a simple discovery dispute. If proven, the allegations would suggest OpenAI deliberately obstructed the court's ability to determine the scope of copyrighted material used in training. Ars Technica characterized the potential misstep as potentially fatal for OpenAI's defense, as courts view spoliation of evidence with severe consequences.
Specific numbers remain contested. The New York Times asserts that OpenAI hid "billions" of logs, while the company has not publicly commented on the figure. The motion seeks sanctions that could range from monetary penalties to an adverse inference instruction, which would allow the jury to assume the hidden evidence hurt OpenAI's case.
The outcome could reshape how AI companies handle training data. A ruling against OpenAI might force greater transparency in machine learning pipelines and set a precedent for copyright holders seeking to prove infringement. TechCrunch notes that news publishers view this as a critical test of accountability.
Legal experts suggest OpenAI faces an uphill battle; concealing evidence in federal litigation often triggers harsh judicial responses.