Adobe has caused controversy over its terms of service updates, with users worried that its works will be used for AI training. Recently, Adobe announced a revision of its terms of service, clearly stating that it will not use customer works for AI training, aiming to calm users' concerns and rebuild trust. The move comes in response to an outcry from users, with Adobe's president admitting the company should have clarified the terms sooner and pledging to be more transparent in the future.
- Adobe revised its terms of service to clarify that it will not use customer works for AI training
- Adobe president admits terms of service should be clarified sooner, says it will be more transparent
- Creators' concerns about Adobe remain, as company works to win back trust
News from ChinaZ.com on June 11: Adobe recently announced that it will revise the terms that users need to agree to when using its applications to make it clear that customer works will not be used for AI training, aiming to win back users. of trust. The change was announced a week ago after an outcry from users.
Users are concerned that an update to Adobe's terms of service will allow the company to freely access and use customers' work to train its generative AI models. However, Adobe President David Wadhwani said that Adobe has never trained generative AI on customer content, has never taken ownership of customer works, and has never allowed access to customer content beyond legal requirements.
Wadhwani acknowledged that the wording in Adobe's terms of service was never meant to allow AI to be trained on customer work, while saying the company should have clarified its terms of service earlier and been more transparent. Previously, Adobe faced widespread criticism from creators who believed that the company had monopolized the industry, adopted a subscription pricing model, and abused generative AI. While Adobe trained its Firefly AI model on Adobe Stock images, some artists have found their work cited on Adobe's platform, making it difficult to believe the protections that exist.
Regarding the content review of Adobe Stock and Firefly training data, Wadhwani said the company has achieved good results, but acknowledged that it will never be perfect. Adobe can remove policy-violating content from Firefly's training data, and customers can opt out of automated systems designed to improve the company's services.
Adobe said in a blog post that they realize "trust must be earned" and are open to feedback to discuss new changes. Greater transparency is a welcome change, but it's likely to take some time to convince harmed creators that the company doesn't have any bad intentions. "We are determined to be a trusted partner for creators in the coming era, and we will work tirelessly to achieve this goal," they said.
Although Adobe has responded, it remains to be seen whether it can completely allay creators' concerns. Earning trust requires continued efforts and more effective communication, and Adobe will still need to invest more in transparency and user rights protection in the future.