Adobe rewords policy after backlash

  • After a poorly worded Terms of Use update saw tempers flare, Adobe has said it will update these terms to be clearer.
  • The company has said that user content belongs to users and it won’t train its AI on that content.
  • However, confidence in the brand has been shaken amidst a growing trend of companies abusing their station to get access to customer data for AI training.

Last week Adobe customers were asked to reaccept Terms of Use and once users started combing through the wording, they found cause for concern. Given the way the policy was worded, users were seemingly agreeing to let Adobe consume their content and use it for training artificial intelligence.

The backlash was severe with many loudly proclaiming they’d be cancelling their Adobe subscriptions and shopping for alternatives. It seems that disdain reached the ears of Adobe’s management team as on Monday the firm announced it would be rewording its terms of use to be less ambiguous about how it uses customer data.

The new Terms of Use will be released next week on 18th June and per the company, will be less ambiguous.

“We’ve never trained generative AI on customer content, taken ownership of a customer’s work, or allowed access to customer content beyond legal requirements. Nor were we considering any of those practices as part of the recent Terms of Use update. That said, we agree that evolving our Terms of Use to reflect our commitments to our community is the right thing to do,” Adobe wrote on Monday.

The company says that the reworded Terms of Use will clarify a number of areas of the policy. These include:

  • You own your content. Your content is yours and will never be used to train any generative AI tool. We will make it clear in the license grant section that any license granted to Adobe to operate its services will not supersede your ownership rights.
  • We don’t train generative AI on customer content. We are adding this statement to our Terms of Use to reassure people that is a legal obligation on Adobe. Adobe Firefly is only trained on a dataset of licensed content with permission, such as Adobe Stock, and public domain content where copyright has expired.
  • You have a choice to not participate in our product improvement program. We may use usage data and content characteristics to improve your product experience and develop features like masking and background removal among others through techniques including machine learning (NOT generative AI). You always have the option of opting out of our desktop product improvement programs.
  • The licenses we require to operate and improve our products on your behalf should be narrowly tailored to the activities needed. The licenses required to operate our products on your behalf use the standard statutory copyright rights but will now include plain English examples of what they mean and why they are required. We will also separate out and further limit the licenses required to improve our products and emphasize the opt-out option. We will reiterate that, in no case do these license grants transfer ownership of your content to Adobe.
  • Adobe does not scan content stored locally on your computer in any way. For content that you upload to our servers — like all content-hosting platforms — Adobe automatically scans content you upload to our services to ensure we are not hosting any child sexual abuse material (CSAM). If our automated system flags an issue, we will conduct a human review to investigate. The only other instances where a human will review your content is upon your request (per a support request) if it is posted to a public facing site, or to otherwise comply with the law.

Adobe acknowledges that as a company that has incorporated AI into its business, it has a responsibility to keep users informed about how their data is used and what it’s used for. Unfortunately, Adobe doesn’t have the best reputation and this latest incident has sent confidence in the company spiralling further downward.

Many Adobe users have said that these forthcoming Terms of Use changes are only coming because Adobe got caught and publicly shamed for trying to push through a policy that practically gave it ownership over a user’s content.

The improper use of user data is currently a major concern with regards to AI platforms. Last month, actress Scarlett Johansson took OpenAI to task for seemingly mimicing her voice for the voice of OpenAI’s Sky. Despite saying it never copied Johansson’s voice, OpenAI paused the rollout of Sky’s voice into its products.

OpenAI itself has openly admitted that without access to up-to-date and copyright protected content, ChatGPT and other solutions would simply be an interesting experiment.

Many users have suggested that Adobe will simply try to sneak through a policy which gives it permission to use user content for training somewhere else in future but we suspect that give this misstep, folks will be reading the terms and conditions a bit more closely.

When it comes to AI, companies have seemingly started to take the approach that it’s better to ask for forgiveness than to ask for permission and users aren’t having any of it.


About Author


Related News