published reporting and published reporting sue OpenAI and Microsoft over AI training

The published reporting and published reporting newspapers have filed a federal lawsuit in the United States against OpenAI and Microsoft, alleging that their journalism was copied without permission to train artificial-intelligence systems. The allegations had not been resolved at the time of the filing, announced on September 5, 2026.

The case adds to the mounting legal pressure around how AI companies obtain their training data. It sharpens scrutiny of data provenance, licensing arrangements and controls over how outputs attribute the material they draw on.

For organisations building or deploying AI, the practical implications are direct: dataset lineage, rights records, retrieval boundaries and litigation holds all need to be maintained. The filing has been independently reported, but the allegations themselves remain unproven, and the case will test where the legal boundaries of AI training on published journalism lie.

A growing number of publishers have taken similar legal action in recent years, arguing that scraping their work without licence undermines the journalism the models draw upon.