For many years, graphic designers considered unapproved use of their work to be an expense of doing business. An unattributed logo or an overly similar style are bothersome, but they are rarely worth the expense of hiring a lawyer. AI is causing that posture to change quickly.
Advertising agencies are being sued by an increasing number of independent and studio-based designers for allegedly feeding client mockups, pitch decks, and portfolios into AI image generators without authorization. The agencies, some of which are among the most prominent names in the industry, are accused of using this scraped material to train internal tools designed to accelerate campaign production. Compared to the lawsuits against Midjourney or Stability AI, this scandal is more subdued, but it is also somewhat more intimate. These are not pictures from the internet that are anonymous. They are designers who submit their work directly to clients, frequently under contracts that make no mention of AI training.
Speaking with those involved in these cases, it’s remarkable how commonplace the betrayal seems to the designers. One illustrator talked about finding her signature linework, created by a tool she had never heard of, replicated in an agency’s internal style guide. No one had hacked her. As she had done a hundred times before, she had just uploaded her file to a shared drive for a campaign review. Designers feel that the trust that is inherent in regular client relationships—submitting your work, waiting for notes, and getting paid—has been subtly abused.
Compared to the major AI lawsuits that are making headlines, these cases are in more ambiguous legal territory. Before a deadline, designers who provide files as part of paid work frequently sign broad usage agreements that no one carefully examines. Agencies are arguing that submission implied consent for internal use, including AI, based on that fine print. The lawyers for the plaintiffs argue that permission to use a design in a campaign does not imply permission to use it as training data for a system that may eventually replace the designer completely. The reason more suits keep coming out is because it’s still unclear which interpretation will hold up.
Beneath all of this is a generational conflict. Senior creative directors, many of whom learned to sketch by hand, appear genuinely concerned about the internal use of AI in their own agencies. In contrast, junior designers are most likely to have their early-career portfolios—work they shared widely in an attempt to gain attention—scraped. Nobody is unaware of the irony: visibility, which was once the main benefit of posting work online, is now a drawback.
The early photography disputes of the 2000s, when stock agencies were accused of underpaying photographers while reselling their images at scale, have drawn comparisons from industry observers. It took more than ten years for that battle to settle into anything approaching fair practice. If only because the financial stakes are higher and the public’s attention is more focused, disputes involving AI scraping might proceed more quickly. However, no one genuinely anticipates a speedy conclusion.
In response, some agencies have quietly updated contracts to include explicit AI-training clauses. Designers view this as both a step forward and proof that something was going on all along. Others have remained silent, which design groups interpreted as a sort of response.
Contracts, rather than courtrooms, will likely determine what happens next. Client agreements can change overnight, but lawsuits take years. As this develops, it’s difficult to ignore the fact that the person who drafts the paperwork first still has a significant amount of power in these disputes. For the time being, designers are advocating for improved disclosure, more precise language, and some control over how their work is used after the invoice is paid. It’s unclear if agencies will confront them there or just wait out the legal commotion.
