Search papers, labs, and topics across Lattice.
This paper investigates the implications of large language models (LLMs) potentially memorizing copyrighted texts, revealing that these models encode statistical relationships between tokens rather than storing information in a traditional database format. The authors argue that current copyright law may classify LLM outputs as "copies" of copyrighted works if they can generate those works with a certain probability, which could have significant legal ramifications for the use of LLMs. Ultimately, they suggest that a functional approach to copyright law may emerge, but they also highlight the need for legal reforms to address these challenges adequately.
LLMs may legally be considered to "copy" copyrighted works if they can probabilistically reproduce them, raising critical questions for AI legality.
Recent work shows that it is possible to extract verbatim or near-verbatim text of some copyrighted works from some large language models (LLMs or models). That is evidence that the model weights encode the works in some form - that the model has"memorized"those works from its training data. But LLMs don't store information in the same format as familiar databases. Rather, their weights store statistical relationships between tokens that have been learned from the training data, and those relationships inform a generation process that is often probabilistic rather than deterministic. In the case of memorization, those relationships are strong enough that, in many circumstances, the model might generate a copyrighted work from its training data with some probability. Copyright law has not previously had to decide whether storing information that might or might not produce output similar to a copyrighted work is itself a copy of the work. The answer to the question is important, because it may determine the legality of many LLMs. The statute and case law are largely unhelpful. We argue that copyright law will likely take a functional approach to the question, finding that LLMs contain a copy of a particular work only if it is straightforward to extract that work in outputs. That result is unsatisfying as a policy matter, and we suggest potential changes to the law, but it is the most likely outcome under current law.
