The problem isn’t Adobe’s fresh terms of service – the problem is trust

A recent announcement from Adobe regarding a terms of service the update caused outrage on the Internet when many people – forced to accept fresh conditions for constant access to applications and services – interpreted it as meaning that Adobe was allowing itself to freely access and employ its work to train artificial intelligence models.

Specifically, the notice noted that Adobe “has made clear that we may access your content through both automated and manual methods” in its Terms of Utilize, directing users to a section that states that “techniques such as machine learning” can be used to analyze content to improve services. software and user experiences. The update went viral after developers took vague language from Adobe to mean it would use their work to train Firefly – the company’s generative artificial intelligence model – or access sensitive projects that may be covered by an NDA.

Adobe has it now published a blog to address these concerns and assure users that this is not the case. “Our commitment to customers has not changed,” Adobe said in a statement, confirming that the company does not train Firefly on customer content or take ownership of customer work. “Firefly’s generative AI models are trained on a dataset that includes licensed content such as Adobe Stock and public domain content where copyright has expired.”

A before and after comparison of the Terms of Use update (which you can see below) shows that not much has actually changed in the rules. In particular, the inclusion of “machine learning”, while vague, is nothing new and has been present in TOS for years. One explanation is that strains of AI technology predating Firefly have long been used in tools like content-aware fill in Photoshop and subject selection in Lightroom. However, what’s confusing is that the updated Terms of Use language has actually been in effect since February, and Adobe only recently notified users of the change.

Adobe Chief Product Officer Scott Belsky confirmed in October said the wording in the notice is “unclear” and said the company’s legal team is working to address concerns over the policy’s unclear language. “Adobe has been offering something like this in TOS for over a decade,” Belsky said. “But trust and transparency couldn’t be more important right now.”

Adobe has created something of an “image” problem that has been growing over the years, especially among individual creators who no longer feel that the company has their best interests at heart. He is criticized for this abandoning the one-time purchase model in favor of recurring subscriptions and accused of creating a monopoly on the creative software industry – which concerned regulators so much that they effectively forced the company to abandon last year’s attempts to take over Figma. Although similar software Is available from other brands such as Affinity, Adobe software is typically considered “industry standard” and difficult to avoid in professional environments.

It’s worth noting that since introducing its own Firefly model in March 2023, Adobe has also developed a whole range of generative AI tools and services, enthusiastically promoting them as a way to quickly create content at scale for people with limited creative experience. I imagine this sounds very attractive to businesses, but not so much to creative professionals who are concerned about their job security. It’s easy to see why so many feel betrayed by the company, especially when it appears to be Adobe he tries enforce generative principles of artificial intelligence introduced to protect them.

This breakdown in trust brings us here. While this viral drama surrounding Adobe’s TOS “update” may be over, legions of creators are watching the company like a hawk. Adobe will need to find an effective way to address these trust issues if it wants independent creators who have come to expect the worst from the company to see it as the friend it claims to be, rather than the enemy.