Meta’s legal strategy for defending its AI training methods came under intense scrutiny this week as a U.S. federal judge questioned the company’s claim that using copyrighted books without permission qualifies as fair use. The lawsuit, filed by authors including Sarah Silverman and Junot Díaz, accuses Meta of copying their works to train its Llama language model without consent or compensation. During Thursday’s hearing in San Francisco, Judge Vince Chhabria challenged the company’s rationale head-on, remarking that generating new content that could directly compete with the original works raises serious questions about market harm. “You’re using protected works to train a system that creates new content which could compete with the originals,” he said. “That sounds like it’s not just eroding the market—it’s potentially destroying it. And you think no compensation is owed?”
- Huawei Developer Conference HDC 2025 live broadcast schedule announced, including keynote speeches, HarmonyOS open classes, etc.
- WeChat supports backup of chat records to USB flash drive and will also support automatic backup – Passionategeekz –
- Huawei MatePad 11.5 2024 tablet opens the Tiansheng Painting App. Public Test
- Huawei’s multiple flagship phones push Hongmeng OS5.0.1.130 optimization gallery and control center – Passionategeekz –
- Huawei nova 12/13 series mobile phones have been upgraded to HarmonyOS 5.0.1.130 public beta version, adding “dynamic photo switch”
Meta contends that its AI models do not reproduce content but learn from it, generating new and creative material that qualifies as transformative—an argument central to U.S. fair use law. However, the judge warned that transformative use alone doesn’t automatically protect against copyright infringement, particularly when there’s clear commercial benefit and potential harm to the original authors’ income.
This clash between Big Tech and creative professionals underscores a growing tension: technology companies argue that strict copyright enforcement could stifle innovation, while authors fear the unchecked use of their work undermines their rights and livelihoods. Meta, like other AI giants including OpenAI and Google, insists that fair use protects their methods and is critical to the future of generative AI development.
Judge Chhabria acknowledged that Meta’s usage might involve some level of transformation, but was unconvinced that this alone made it legally permissible. The case could set a precedent not only for Meta but for the entire AI industry, determining whether training models on copyrighted material without permission will remain standard practice or become a costly legal misstep.
While no ruling has been issued yet, the message from the bench was clear: the fair use defense will not go unchallenged. As generative AI reshapes the boundaries of authorship and originality, the courts are preparing to define just how far these models can go without stepping over the legal line.
[via]
Discover more from PassionateGeekz
Subscribe to get the latest posts sent to your email.