正文：
近日，Adobe公司因其AI图像生成器Firefly的训练数据来源问题而陷入了舆论的漩涡。据彭博社报道，Adobe在训练Firefly时，不仅使用了自家库存数据库中的图像，还秘密地使用了来自竞争对手Midjourney以及其他图像AI的图像。这一行为与Adobe此前对外宣称的仅使用自有或公共领域内容训练Firefly的说法相悖。
Adobe方面对此回应称，他们确实从第三方AI中获取了图像，但这些图像占训练数据总量的5%。Adobe强调，这部分图像的法律和道德背景并不明确，因此在公开声明中并未提及。Adobe表示，他们使用这些图像是为了确保Firefly的图像质量，使其更加逼真。
这一消息引起了业界和公众的广泛关注和讨论。一方面，有人认为Adobe的行为是业界竞争中的正常现象，AI技术的进步往往需要大量的数据来训练模型。另一方面，也有人质疑Adobe的透明度，认为公司在处理第三方数据时应该更加谨慎，确保不会侵犯他人的版权或隐私。
Adobe的这一做法也引发了对于AI训练数据来源的讨论。业界专家指出，随着AI技术的快速发展，如何合法合规地获取和利用数据将成为企业必须面对的挑战。同时，这也提醒了消费者和监管机构需要加强对AI技术的监管，确保技术的健康发展。
目前，Adobe已经表示将就此问题进行内部审查，并承诺在未来会提供更加透明的数据来源信息。Adobe的这一事件再次提醒了业界，在AI技术的快速发展过程中，透明度和合规性是企业必须坚守的底线。
英语如下：
News Title: “Adobe Firefly AI Suspected of Learning from Rival Midjourney”
Keywords: Adobe, Image Generation, Competitor Training
News Content:
Title: Adobe’s Use of Competitor Images in Training AI Generator Sparks Controversy
Article:
Recently, Adobe Inc. has found itself at the center of a media storm over the source of training data used for its AI image generator, Firefly. According to Bloomberg, while training Firefly, Adobe not only utilized images from its own stock database but also secretly employed images from rivals such as Midjourney and other AI image generators. This practice contradicts Adobe’s previous public statement that it only trains Firefly with content from its own databases or the public domain.
Adobe responded to the allegations by acknowledging that they did obtain images from third-party AI sources, but claimed that these images only constitute 5% of the total training data. Adobe emphasized that the legal and ethical background of these images was not clear, hence they were not mentioned in public statements. The company stated that they used these images to ensure the high quality of the images produced by Firefly, making them more realistic.
The news has sparked widespread attention and debate within the industry and among the public. On one hand, some argue that Adobe’s actions are a normal part of industry competition, as the advancement of AI technology often requires vast amounts of data to train models. On the other hand, there are concerns about Adobe’s transparency, with some questioning why the company should be so cautious in handling third-party data to avoid infringing on copyright or privacy.
Adobe’s actions have also raised discussions about the sources of AI training data. Industry experts point out that as AI technology continues to develop rapidly, the legal and ethical considerations of obtaining and utilizing data will become a challenge that companies must address. This situation also reminds consumers and regulatory bodies of the need to strengthen oversight of AI technology to ensure its healthy development.
In response, Adobe has stated that it will conduct an internal review of the matter and commit to providing more transparent information about the data sources in the future. Adobe’s incident serves as a reminder to the industry that transparency and compliance must be upheld as the industry continues to develop rapidly.
【来源】https://the-decoder.com/adobe-trained-its-ai-image-generator-on-midjourney-images-but-its-complicated/
Views: 6