美国众议员 Adam Schiff 近日提出了一项名为《生成式 AI 版权披露法案》的新立法，旨在加强人工智能（AI）行业对版权保护的透明度。根据 Billboard 的报道，该法案要求AI公司公开其模型训练过程中使用了哪些受版权保护的数据，否则将面临经济处罚。
该法案不仅针对未来开发的生成式 AI 模型，还将溯及已发布的、正在运行的AI系统。法案规定，AI公司在其模型向公众提供前的30天内，必须向美国版权局提交一份详尽的清单，列出所有在训练过程中使用的受版权保护作品。此外，一旦这些模型的训练数据集发生“重大改变”，AI公司也需及时更新并向版权局报告。
这一举措反映出立法者对保护知识产权和防止潜在版权侵权问题的关注，旨在为AI技术的发展设定清晰的法律边界。如果该法案获得通过，将对AI行业产生深远影响，促使企业更加谨慎地处理训练数据，尊重并保护原创作品的版权。
英语如下：
News Title: “US Lawmaker Proposes Bill: AI Companies Must Disclose Copyrighted Training Data or Face Penalties”
Keywords: AI Bill, Copyright Disclosure, Economic Penalties
News Content: U.S. Representative Adam Schiff recently introduced a new legislative proposal called the “Generative AI Copyright Disclosure Act,” aimed at enhancing transparency in the artificial intelligence (AI) industry regarding copyright protection. According to Billboard, the bill would require AI companies to disclose which copyrighted materials were used in their models’ training processes, with financial penalties for non-compliance.
The bill not only applies to future generative AI models but also extends to already deployed and operational AI systems. It stipulates that AI companies must submit a comprehensive list to the U.S. Copyright Office 30 days prior to making their models available to the public, detailing all copyrighted works used during the training process. Furthermore, should there be “significant changes” to the training dataset, companies would be required to promptly update and report these to the copyright office.
This move reflects lawmakers’ concern for protecting intellectual property and mitigating potential copyright infringement issues, seeking to establish clear legal boundaries for the development of AI technology. If passed, the bill would have a profound impact on the AI industry, encouraging companies to handle training data more cautiously and respect the copyrights of original works.
【来源】https://www.ithome.com/0/761/336.htm
Views: 1