Ars Technica reports that The New York Times has filed a heavily redacted court document accusing the software king of the world, Microsoft, of actively encouraging OpenAI to nick its copyrighted work.
The move follows the US Supreme Court’s Cox Communications ruling, in which Sony failed to hold an internet service provider liable for contributory music piracy.
That ruling shifted the legal bar, meaning plaintiffs must show that a defendant intentionally induced unlawful conduct.
The Times wants to tweak its complaint so its contributory infringement claim against Vole fits that nastier new standard.
A Volish spokesperson called the amended complaint “a last-ditch effort by the plaintiff to save its claim from unfavourable precedent set in other recent rulings…”
The updated complaint claims Microsoft’s supercomputer was designed for OpenAI with one purpose in mind: training AI on copyrighted work without permission.
The New York Times claims the system gave its journalism extra weight because both firms wanted models that could mimic top-tier reporting.
“Microsoft designed it for the purpose of using the whole Internet essentially — curated to disproportionately feature Times Works — to train the most capable LLM in history,” the New York Times alleged.
The paper said Microsoft did more than offer plumbing. It allegedly helped select the works being infringed and supplied the means to grab them. The newspaper is narked about hallucinations too, especially when Microsoft and OpenAI models falsely cite it for articles it never published.
“Users who ask a search engine what The Times has written on a subject should be provided with neither an unauthorised copy nor an inaccurate forgery of a Times article, but a link to the article itself,” the New York Times alleged.
OpenAI spokesperson Drew Pusateri said the outfit still believes training AI on copyrighted work is fair use. OpenAI has argued that ChatGPT does not replace a Times subscription because the system transforms material for a different use.
An OpenAI spokesperson told Ars Technica that its models “empower innovation,” which is the sort of phrase corporate lawyers keep in a drawer.
The New York Times is not buying it.
“Our core claims remain the same from the day we filed this lawsuit — that Microsoft and OpenAI stole millions of The Times’s copyrighted works to compete with our products and illegally enrich themselves,” a New York Times spokesperson said.
The nightmare scenario for Microsoft and OpenAI would be a court order forcing them to wipe trained models and start again.
