2026-11-04 · Fonte:ZDNet

The DOJ calls AI training transformative fair use; publishers warn unlicensed scraping threatens their survival in a fight that could shape online knowledge.
The US Department of Justice has placed itself at the center of the AI industry's most consequential legal fight, arguing that training artificial-intelligence models on copyrighted material amounts to transformative fair use. ZDNet frames the stakes more starkly, contending in its headline that the DOJ's OpenAI stance is "the real threat to national security" — recasting a copyright dispute as a question about the foundations of American information.
Against that position, publishers argue that unlicensed scraping — the wholesale ingestion of their work without authorization — threatens their survival. Their case rests on a simple premise: professional publishing cannot persist if the content it produces can be freely absorbed into commercial AI systems. In this reading, fair use for training is not a technical legal doctrine but a decision about whether the people who create knowledge can keep funding its creation.
The significance reaches beyond any single case. As the report stresses, this copyright fight could shape the future of online knowledge itself. The two positions point in opposite directions: one treats training as a transformative act that the law should permit, the other describes a practice eroding the publishers who produce much of what the web has to offer. Whichever interpretation prevails will influence what gets published, how it is protected, and who captures its value.
For an industry accustomed to measuring progress in chips and data centers, the dispute is a reminder that model-building depends on a legal input as well: the permissibility of training data. With the DOJ endorsing transformative fair use and publishers warning that unlicensed scraping undermines their survival, the outcome of this fight will help set the terms on which the next generation of AI systems — and the knowledge they draw on — gets built.
Tags: OpenAI · DOJ · Copyright · Fair Use · AI Training
萬安算交所 AIXX · AIXX · Exchange verificado · Ir ao mercado