据彭博社近日报道，全球知名创意软件巨头Adobe在训练其AI图像生成器Firefly的过程中，采用了竞争对手Midjourney等公司的图像数据，这一行为与其此前的公开声明存在出入。Adobe一直宣称，Firefly的训练仅基于公司拥有权利或公共领域内的内容，强调使用自身的库存数据库作为训练材料。然而，The Decoder的最新报道揭示，实际情况并非如此简单。
据透露，Adobe在技术上确实使用了自家库存数据库中的材料训练Firefly，但其中竟有5%的训练数据来自来源不明的人工智能图像，这5%的数据可能涉及法律和道德上的争议。这一比例虽然不大，但足以引发对其透明度和合规性的质疑。
这一新闻瞬间引发了业界的广泛关注，因为Adobe作为行业领导者，其在数据使用和AI伦理方面的做法具有广泛的影响力。如果Adobe未能充分披露其训练AI模型的具体细节，可能导致公众对其信任度的下降，并可能引发关于AI开发标准和知识产权的广泛讨论。
目前，Adobe尚未就这一报道发表官方回应。然而，这一事件无疑提醒了科技行业，透明度和责任在AI发展中的重要性，特别是在涉及第三方数据和潜在法律风险时。随着AI技术的快速发展，企业如何在利用数据的同时确保合规和透明，将成为未来面临的关键挑战。
英语如下：
News Title: “Adobe Firefly Training Controversy: Alleged Use of Rival Midjourney Images Raises Transparency Questions”
Keywords: Adobe, Firefly, Image Training
News Content:
Title: Adobe Firefly’s Training Process Under Scrutiny Amid Reports of Using Rival Midjourney’s Image Data
According to recent reports by Bloomberg, global creative software giant Adobe is alleged to have utilized image data from competitors like Midjourney during the training of its AI image generator, Firefly, contradicting previous public statements. Adobe has consistently claimed that Firefly’s training is based solely on content for which it holds rights or is in the public domain, emphasizing the use of its own stock database as training material. However, new revelations by The Decoder suggest a more complex situation.
It has been disclosed that Adobe did indeed use material from its own stock database to train Firefly, but a surprising 5% of the training data came from unattributed AI-generated images, potentially involving legal and ethical concerns. While this percentage is small, it has raised questions about transparency and compliance.
This news has immediately drawn widespread attention in the industry, as Adobe, as a leader, carries significant influence in data usage and AI ethics. If Adobe fails to fully disclose the specifics of its AI model training, it could undermine public trust and spark broader discussions on AI development standards and intellectual property.
As of now, Adobe has not issued an official response to these allegations. Nevertheless, this incident underscores the importance of transparency and accountability in AI development, especially when dealing with third-party data and potential legal risks. As AI technology advances rapidly, how companies balance data utilization with compliance and transparency will be a crucial challenge in the future.
【来源】https://the-decoder.com/adobe-trained-its-ai-image-generator-on-midjourney-images-but-its-complicated/
Views: 3