
Artist style and copyright: Adobe's legal stance on AI content
Adobe's deputy general counsel has outlined the company's position on protecting artists' style under copyright, in the US and Europe. The report gives no detail on the substance of that position or on the legal advice offered to people creating with AI.
Blog du Modérateur reports that Karen Robinson, Adobe's Deputy General Counsel, has set out the publisher's position on protecting artists' style, both in the United States and in Europe, and offered legal guidance to those who create with AI. The report does not specify what that position contains, which legal mechanisms it relies on, or what the advice actually recommends.
That silence is the story worth noting. A company that both sells creative tools and trains or licenses models built on creative work has to draw a line somewhere. The title of the report suggests where that line falls: established characters and franchises are treated as protected, while an individual artist's style is not. If that reading holds, it matches how copyright has long worked in practice.
Style was never the protected asset
Copyright protects expression fixed in a tangible form, not ideas, methods or aesthetic directions. A distinctive way of drawing, a palette, a way of composing a frame: these have historically been hard to own, and courts in several jurisdictions have been reluctant to treat style itself as property. What is protected is a specific work, and in some countries moral rights that attach to it.
Context matters here. The wider debate over generative models has focused on training data and on outputs that closely reproduce protected works, not on style as such. The report gives no figures, no case references and no jurisdictions beyond the US and Europe, so any conclusion about enforcement remains open.
What a creative or marketing team can check
The practical question for teams producing content with AI is not whether a style can be owned, but whether a given output is defensible. The report offers no checklist, so the following is general practice rather than guidance drawn from it.
- Keep a record of which tool produced which asset, under which plan and terms of service, and when.
- Review the commercial terms of each model you use: some restrict use of outputs, some offer indemnification, most differ by plan.
- For anything featuring a recognizable character, brand or living person, treat the asset as high-risk regardless of the tool.
- Ask whether a human contribution (selection, editing, composition) is documented, since that record matters if authorship is ever contested.
The gap that remains
A vendor's public position is not a legal ruling. Until courts or legislators in the US and Europe settle how training and outputs are treated, the safe assumption is that responsibility sits with whoever publishes the content. The report does not say whether Adobe's advice addresses that allocation, and it does not indicate any timeline for regulatory clarity.
The useful move this week is administrative: map which AI tools touch your published assets, confirm what each contract says about outputs, and flag any asset built on a recognisable character or a living artist's signature look. Style may not be ownable, but the commercial risk around it is real.
Also worth reading
-
LinkedIn shares tips on getting cited by AI models
LinkedIn has published guidance on how companies can improve their chances of being cited by large language models. The report points to credibility and the authority of influencers as key factors, though the underlying mechanics remain undisclosed.
-
AI at work: study of 9,000 employees finds no clear time savings
An ActivTrak study of nearly 9,000 employees over six months challenges the idea that AI tools automatically save working time. For marketing and growth teams, it raises a practical question about where AI actually moves the needle.
-
More AI tools, less growth: what CMOs are getting wrong
A Digiday piece reports that enterprise CMOs see no growth from added AI spend, only task-level efficiency. The gap comes from siloed pilots and delegated ownership rather than from the models themselves.


