A US federal judge dismisses Raw Story and AlterNet's copyright lawsuit against OpenAI over LLM training, for now, saying the outlets couldn't show enough harm
A New York federal judge on Thursday dismissed a lawsuit against artificial intelligence giant OpenAI that claimed it misused articles …
Context & Ripple Effects
Raw Story and AlterNet joined the early wave of publisher challenges to OpenAI when they sued alongside a separate Intercept case in New York. Their claim now diverges procedurally from the broader publisher-led challenge filed in February 2024 because the court found the alleged harm insufficient at this stage.
The ruling matters because other training-data cases have turned on which legal theories and factual allegations survive early dismissal. In related coverage, a California court also allowed an unfair-competition claim over copyrighted-book training to proceed while dismissing other counts.
First-order effects
- Raw Story and AlterNet's case against OpenAI is dismissed for now, removing its immediate litigation pressure unless the outlets can remedy the court's finding on harm.
- OpenAI avoids having to defend this particular complaint on the merits at this stage; the decision does not resolve the wider question of whether LLM training on news articles is lawful.
Second-order effects
- Publishers pursuing training-data claims will face stronger incentives to tie alleged use of their work to concrete, legally cognizable harm rather than relying on the act of training alone.
- The contrast with cases that preserve at least some claims, including the copyrighted-books unfair-competition claim, makes pleading strategy and jurisdiction more consequential across AI-content litigation.
Third-order effects
- If courts continue to sort these suits primarily by demonstrated harm and claim construction, copyright litigation may produce uneven precedents rather than a single early rule for LLM training.
- That uncertainty keeps commercial arrangements for AI content under pressure: rights holders gain leverage where they can document harm, while model developers retain room where claims fail at the pleading stage.
The trend: AI training-data disputes are evolving into a case-by-case test of whether rightsholders can convert alleged unauthorized use into provable legal harm.