据《The Decoder》报道，全球知名软件公司Adobe近期被曝出在训练其AI图像生成器Firefly的过程中，使用了竞争对手Midjourney等公司的图像数据。这一消息与Adobe早前的声明相悖，该公司曾表示，Firefly的训练仅基于其拥有版权或属于公共领域的内容。
据彭博社的进一步报道，Adobe确实遵循了技术上的准确，即使用自家库存数据库中的材料对Firefly进行训练。然而，一个未被广泛披露的细节是，这些训练数据中竟然有5%的来源是存在法律和道德争议的人工智能生成图像。这一情况使得Adobe在透明度方面受到了质疑。
Adobe尚未对这一新的指控发表官方回应，但这一发现引发了业界对于AI开发中数据来源和透明度的讨论。如果属实，Adobe可能需要重新审视其数据使用策略，以确保其在创新的同时，遵守版权法规并保持诚信。这一事件也提醒了公众和行业，对于AI技术的发展，不仅需要关注技术的进步，更应关注其背后可能涉及的法律和伦理问题。
英语如下：
News Title: “Adobe Firefly Allegedly Trained with Rival Midjourney’s AI Images, Sparking Transparency Concerns”
Keywords: Adobe, Firefly, AI Training
News Content: According to a report by “The Decoder,” renowned software company Adobe has been revealed to have utilized image data from competitors like Midjourney to train its AI image generator, Firefly. This contradicts Adobe’s earlier statements, which claimed that Firefly’s training was exclusively based on content for which it held copyrights or was in the public domain.
Further reporting by Bloomberg disclosed that Adobe did indeed train Firefly using materials from its own stock database, as technically claimed. However, an underreported detail is that 5% of the training data came from AI-generated images with contentious legal and ethical implications. This has raised questions about Adobe’s transparency.
Adobe has yet to issue an official response to these new allegations. The revelation has sparked industry discussions about the sourcing of data and transparency in AI development. If true, the company might need to reassess its data usage strategy to ensure compliance with copyright laws and maintain integrity. This incident serves as a reminder to the public and the industry that, alongside technological advancements, attention must also be given to the legal and ethical issues that AI development may entail.
【来源】https://the-decoder.com/adobe-trained-its-ai-image-generator-on-midjourney-images-but-its-complicated/
Views: 1