Intellectual property rights

Mitigation strategy

1. Establish a comprehensive Intellectual Property (IP) governance framework, including internal AI use policies that mandate strict ethical and legal data sourcing procedures for all training datasets. This must prioritize obtaining explicit licenses from copyright and database right owners or rigorously ensuring the lawful application of available exceptions, such as Text and Data Mining (TDM) provisions and rightsholder opt-out mechanisms. 2. Implement a robust, multilayered pre-training data filtering pipeline—incorporating content verification, machine learning classifiers, and continuous database cross-referencing—to shift copyright protection from post-training detection to proactive, technical prevention of unauthorized content ingestion. 3. Conduct rigorous IP due diligence on all third-party data sources and AI vendors. This includes securing enterprise-grade licenses with clear indemnification clauses and defined IP ownership terms to mitigate risks arising from third-party infringement claims related to the model's training data.