Position: No Retroactive Cure for Infringement during Training
For the machine learning community facing legal challenges, this paper clarifies that post-hoc methods are legally insufficient, urging a focus on data provenance during training.
This paper argues that post-hoc mitigation methods like machine unlearning cannot retroactively cure infringement liability from unauthorized data acquisition and training, because compliance depends on data lineage, not outputs. The authors advocate for a shift from post-hoc sanitization to verifiable ex-ante process compliance.
As generative AI faces intensifying legal challenges, the machine learning community has increasingly relied on post-hoc mitigation -- especially machine unlearning and inference-time guardrails -- to argue for compliance. This paper argues that such post-hoc mitigation methods cannot retroactively cure liability from unlawful acquisition and training, because compliance hinges on data lineage, not the outputs. Our argument has three parts. First, unauthorized copying/ingestion can be a legally complete completed act, and model weights may operate as fixed copies that retain training-derived expressive value, making later filtering beside the point for infringement. Second, contract and tort/unfair-competition rules -- via licenses, terms of service, and anti-free-riding principles -- can independently restrict access and use, often bypassing copyright defenses (e.g., fair use or TDM exceptions). Third, since value from protected inputs can persist in weights, remedies such as unjust enrichment and disgorgement may require stripping gains and, in some cases, reaching the model itself. We therefore argue for a shift from Post-Hoc Sanitization to verifiable Ex-Ante Process Compliance.