Big AI's content problem: Take the work, keep the money (www.theregister.com)

Big AI is facing a significant legal challenge over the practices of utilizing content generated by others—such as books, articles, and code—to train large language models (LLMs). In a court case involving allegations of copyright infringement, Microsoft’s Dr. Brent Hecht bluntly stated that this represents "the largest theft of labor in human history." The case, with ongoing deliberations by Judge Sidney H. Stein, centers around whether the use of public data for training is a legitimate “fair use,” or if it undermines the very creators from whom this content is sourced. Key statements from OpenAI's leadership indicate an awareness of potential legal liabilities, as they have reportedly trained their models on content even behind paywalls without ensuring compliance with copyright protections. This situation is significant for the AI/ML community as it highlights a fundamental tension between innovation and intellectual property rights, raising questions about the sustainability of current business practices in Big AI. The implications of this legal discourse are profound; if successful, it could lead to stricter requirements for content usage in AI training, directly affecting how AI models are developed. Additionally, it underscores a broader concern regarding the economic viability of creators in light of AI systems that could potentially erode their livelihoods while establishing monopolistic control over the data supply chain, compelling a re-evaluation of how AI companies engage with original content produced by individuals.