Copyright in AI Pre-Training Data Filtering: Regulatory Landscape and Mitigation Strategies
Mariia Kyrychenko, Mykyta Mudryi, Markiyan Chaklosh
TL;DR
The paper analyzes how regulatory frameworks across the EU, US, and Asia-Pacific shape AI training data governance and identifies critical gaps that undermine creator rights and AI development. It argues that current enforcement is largely reactive, with two fundamental challenges: the impracticality of comprehensive pre-training license collection and the absence of verifiable verification mechanisms. A multilayered filtering pipeline is proposed, combining access control, pre-crawl verification, perceptual hashing, named entity recognition, tailored classifiers, and continuous database cross-referencing, underpinned by transparent data provenance. If implemented with strong regulatory mandates, this defense-in-depth approach could shift copyright protection from post-hoc detection to pre-training prevention, enabling safer and more responsible GPAI progress.
Abstract
The rapid advancement of general-purpose AI models has increased concerns about copyright infringement in training data, yet current regulatory frameworks remain predominantly reactive rather than proactive. This paper examines the regulatory landscape of AI training data governance in major jurisdictions, including the EU, the United States, and the Asia-Pacific region. It also identifies critical gaps in enforcement mechanisms that threaten both creator rights and the sustainability of AI development. Through analysis of major cases we identified critical gaps in pre-training data filtering. Existing solutions such as transparency tools, perceptual hashing, and access control mechanisms address only specific aspects of the problem and cannot prevent initial copyright violations. We identify two fundamental challenges: pre-training license collection and content filtering, which faces the impossibility of comprehensive copyright management at scale, and verification mechanisms, which lack tools to confirm filtering prevented infringement. We propose a multilayered filtering pipeline that combines access control, content verification, machine learning classifiers, and continuous database cross-referencing to shift copyright protection from post-training detection to pre-training prevention. This approach offers a pathway toward protecting creator rights while enabling continued AI innovation.