STM Association publishes a formal discussion document calling on GenAI developers to treat scholarly content differently from general web content — acknowledging the unique properties of peer-reviewed, versioned, integrity-assured research. The document explicitly invites responses from AI companies, making this an active policy move rather than just commentary. This is a significant industry coordination moment: the major publisher trade body staking out principles for how AI should handle research content, with direct implications for training data rights, attribution, and the treatment of retractions and corrections in AI outputs. Top priority for p2preview.

Source: STM Publishing News