Online legal database CanLII sues Caseway AI for copyright infringement

The legal dispute between the Canadian Legal Information Institute and AI startup Caseway AI highlights the growing tension between open legal data providers and artificial intelligence companies seeking training materials. CanLII alleges that Caseway illegally scraped millions of court records, violating terms of use and infringing copyright, while Caseway denies any wrongdoing, claiming it only uses public records and disputes the legal validity of binding website terms. This conflict represents a critical battleground for defining the boundaries of data access, as non-profits invest significant resources to curate public information while startups argue for unrestricted access to public domain works. The case raises fundamental questions about whether enhancing public records with metadata or structure creates protectable intellectual property or if such data remains freely available for AI analysis. Legal experts suggest the outcome will depend on whether CanLII’s editorial choices meet the threshold for originality, a determination that could set precedents for how AI models interact with digitized legal archives. Furthermore, the lawsuit underscores the practical difficulties smaller AI innovators face in sourcing high-quality data, as major publishers increasingly restrict crawling, potentially stifling competition and innovation in the legal tech sector. This litigation is highly relevant to open data advocates because it tests the resilience of public access principles in the age of generative AI. If CanLII succeeds, it may encourage other non-profit data custodians to restrict bulk access, limiting the ability of new models to train on comprehensive datasets. Conversely, a ruling in favor of Caseway could establish that public legal records remain free for AI ingestion, preserving the openness essential for transparent and equitable legal research. The outcome will likely influence global policies on data scraping, copyright, and the sustainability of public interest information repositories.

Source: betakit.com
Published on 2024-11-09