【新闻报道】近日，一项旨在规范人工智能领域版权使用的全新计划——Fairly Trained 正式启动，由曾担任 Stability AI 音频副总裁的 Ed Newton-Rex 策划并创立。Fairly Trained 计划致力于为人工智能模型公司提供权威认证服务，确保他们在开发过程中合法合规地使用受版权保护的训练数据。
该计划近期推出的首个认证项目名为“许可模型认证”（Licensed Model certification），专为那些主动申请并获得使用版权受限训练数据许可的AI模型公司颁发。这一举措对于促进人工智能产业健康发展具有重要意义，表明了企业对于尊重原创、合法使用知识产权的高度责任感。
截至目前，Fairly Trained 已经完成了对九家在图像生成、音乐创作以及语音合成等领域具备先进技术实力的生成式人工智能公司的认证工作，其中包括那些通过获取授权成功运用版权受保护训练数据的企业。这些被认证的公司在未来的产品和服务中，将显著标注“许可模型认证”标识，以向用户传达其在版权合规性上的严格遵循与承诺。
此举无疑为全球范围内的人工智能行业树立了一个积极范例，推动了业界对于版权保护和合理使用数据的认知提升，并为构建更加公平、透明和负责任的人工智能生态贡献了重要力量。
英语如下：
Headline: “Fairly Trained Innovation Practice: First ‘Licensed Model Certification’ Awarded as AI Industry Steps into a New Era of Legally Copyright Data Usage!”
Keywords: Fairly Trained, Licensed Model Certification, AI copyright authorization
News Content: 【Press Release】Recently, a groundbreaking initiative aimed at standardizing the use of copyright in the artificial intelligence (AI) sector – Fairly Trained – has officially launched. The project was conceptualized and established by Ed Newton-Rex, who served as Audio Vice President at Stability AI. Fairly Trained is committed to providing authoritative certification services for AI model companies, ensuring their compliance with copyright laws during the development process when utilizing copyrighted training data.
The first certification program rolled out under this initiative is called “Licensed Model Certification,” specifically designed for AI model companies that proactively apply and obtain permissions to use copyrighted, restricted training data. This move holds significant implications for fostering the healthy growth of the AI industry, reflecting the high level of corporate responsibility these entities exhibit towards respecting original works and legally utilizing intellectual property rights.
As of now, Fairly Trained has completed certification for nine cutting-edge generative AI companies operating in domains such as image generation, music composition, and speech synthesis, including those that have successfully employed licensed copyrighted training data. These certified companies will prominently label their products and services with the “Licensed Model Certification” seal, underscoring their strict adherence to and commitment to copyright compliance.
This development undoubtedly sets a positive example globally for the AI industry, raising awareness about copyright protection and responsible data usage. It contributes significantly to the establishment of a fairer, more transparent, and accountable AI ecosystem.
【来源】https://www.theverge.com/2024/1/17/24041518/generative-ai-copyright-violation-fair-training-label-certification
Views: 1