近日，由前 Stability AI 音频副总裁 Ed Newton-Rex 创立的 Fairly Trained 计划，针对未侵犯版权的人工智能模型颁发认证。该计划旨在为 AI 模型公司颁发/标记一个认证，以证明他们已申请使用受版权保护的训练数据的许可。
首个认证项目名为“许可模型认证”（Licensed Model certification），将颁发给那些许可受保护数据来训练其模型的公司。目前，已有 9 家从事图像、音乐和语音生成业务的生成式人工智能公司获得了这一认证。
Fairly Trained 计划的实施，意味着人工智能模型公司在使用受版权保护的数据时，将更加规范和合法。这不仅有助于保护版权持有者的权益，也有利于人工智能模型公司的可持续发展。此外，这一认证也能提高用户对人工智能模型的信任度，推动人工智能行业的健康发展。
Title: Certification Program for Artificial Intelligence Models Launched
Keywords: AI models, copyright certification, Licensed Model certification
News content: Recently, the Fairly Trained program, founded by Ed Newton-Rex, former vice president of audio at Stability AI, has started to issue certifications for artificial intelligence models that have not infringed on copyrights. The program aims to award/mark a certification to AI model companies, proving that they have applied for a license to use copyrighted training data.
The first certification program is called “Licensed Model certification,” which will be issued to companies that have obtained permission to use protected data to train their models. Currently, nine generative AI companies involved in image, music, and speech generation have received this certification.
The implementation of the Fairly Trained program means that AI model companies will be more standardized and legal when using copyrighted data. This not only helps protect the rights of copyright holders but also promotes the sustainable development of AI model companies. Furthermore, this certification can enhance user trust in AI models and contribute to the healthy development of the artificial intelligence industry.
【来源】https://www.theverge.com/2024/1/17/24041518/generative-ai-copyright-violation-fair-training-label-certification
Views: 1