NYT slams Microsoft for building copyright-infringing supercomputer for OpenAI

Key Points:

• The New York Times (NYT) has filed a heavily redacted motion to amend its copyright lawsuit against OpenAI and Microsoft, alleging Microsoft actively encouraged OpenAI to steal NYT content by building a custom supercomputer designed for training AI on copyrighted works without permission.
• This amendment follows a recent Supreme Court ruling raising the standard for contributory infringement, requiring plaintiffs to prove intentional inducement of illegal conduct; NYT aims to align its claims with this new legal precedent.
• NYT asserts that Microsoft’s bespoke supercomputing system disproportionately used NYT articles to train AI models, resulting in market harm by substituting NYT subscriptions and reputational harm from AI-generated false attributions.
• Evidence presented by NYT includes ChatGPT outputs showing near-verbatim reproductions of NYT articles and instances where AI falsely cited or fabricated NYT content, which NYT argues demonstrates direct market substitution and damages.
• OpenAI maintains that its AI training constitutes fair use and that ChatGPT is not a substitute for a NYT subscription, but NYT seeks permanent injunctive relief and damages, warning that if the court sides with them, OpenAI and Microsoft may be forced to remove infringing content from their models.