NYT slams Microsoft for building copyright-infringing supercomputer for OpenAI

According to Ars Technica AI, The New York Times has escalated its legal battle against Microsoft over the construction of a supercomputer designed to train OpenAI's models, alleging that the system was built using copyrighted journalistic content without proper licensing. The NYT's argument pivots on a recent Supreme Court ruling against Sony, which the paper claims sets a precedent that Microsoft's infrastructure facilitated copyright infringement at scale. This development adds a new dimension to ongoing lawsuits over AI training data, shifting focus from OpenAI's outputs to the hardware and software pipeline that powers them. For developers and solopreneurs building AI workflows, the case highlights the increasing legal scrutiny of the entire AI supply chain—from data sourcing to model training. It underscores the importance of verifying that any third-party models or training datasets used in production do not expose the builder to downstream liability. The outcome could influence how AI tools are licensed and deployed, potentially raising costs and compliance requirements for those relying on foundation models trained on publicly scraped data.