The proliferation of artificial intelligence image and video generators has made it easy for online users to create billions of memes ranging from baby versions of The Real Housewives of Atlanta to humorous takes on the Coldplay kiss cam viral moment just a few days ago.
But these tools have raised serious legal questions about the copyright protection for the assets that these AI models are trained on.
That has resulted in numerous lawsuits being filed by individual artists, Hollywood studios, and media companies, who all assert that some of the most popular AI systems are trained on unauthorized images and videos. Ongoing litigation includes Disney and Universal suing image generation tool Midjourney, the New York Times squaring up against ChatGPT owner OpenAI and Microsoft, and the Wall Street Journal and New York Post versus AI startup Perplexity.
Ely Greenfield, chief technology officer at software maker Adobe’s digital media business, has spent over two years pitching a different path. Ever since the debut of a text-to-image model known as Firefly in March 2023, Adobe has touted the company’s own creative generative AI models that are only trained on content that it has rights to use, including Adobe Stock photos and licensed artistic content.
Firefly’s models have been integrated in Adobe’s suite of apps including Photoshop and Illustrator, and thus far, businesses and individual creators have generated over 26 billion assets. Big names including toy maker Mattel and cosmetics manufacturer Estée Lauder have signed on to Firefly for creative ideation, editing, and asset generation purposes.
“Every piece of content that we train on is something that we have acquired the license of, or that is published under a verifiable and known license,” says Greenfield.
This approach does come with some limitations. If Firefly were asked to generate an image of a Disney cartoon character, like say Mickey Mouse, “it would do a horrible job of it,” concedes Greenfield. “And that’s by design and on purpose.”
Greenfield says that AI tools based on every image found on the internet produce less desired outputs, not just for potentially infringing on IP, but because it is representative of a vast trove of data that doesn’t always have the best quality. “There’s the raw science of how you build the model, but a massive amount of work goes into data curation and preparation,” says Greenfield. “The average piece of content on the internet isn’t necessarily what you want to put in your ad.”
Adobe’s buttoned-up AI approach means the company’s off-the-shelf Firefly offering would have little use to a consumer-facing company like Coca-Cola. But under an enterprise licensing agreement, Adobe says it can train a private version of Firefly that’s exclusively trained on the beverage company’s branding and style.
Since Firefly’s launch, Adobe has had to make some modifications to the images in the company’s asset bank. Early on, generative AI wasn’t great at producing clear images of hands, so Adobe had to reach out to the photographers it works with to get more licensed pictures of hands to train the AI properly.
All Firefly content also goes through a moderation process that includes a mix of human and computer oversight, eliminating harmful images, but also those that may contain sensitive IP. A photographer may have exclusive license to an image that they produced, but if there’s a trademark asset like a Nike Swoosh or Starbucks Siren logo, Adobe will nix the image.
Adobe has lauded the proliferation of Firefly, reporting in the most recent second fiscal quarter ending May 30 that traffic to the Firefly App grew 30% from the prior quarter, with paid subscriptions nearly doubling over the same period.
More recently, Adobe has integrated image and video models from OpenAI, Google, Pika Luma AI, and Runway into the company’s Firefly app.
This runs parallel with the public’s shifting views on the ethical uses of AI, as well as some recent court decisions that AI hyperscalers have won. Anthropic, in one example, saw a ruling go its way last month that said the company could train models using published books without consent from the authors. To be sure, it will be years before the courts resolve these thorny legal matters, and the right use of images, text, and audio assets will almost certainly vary across the globe.
For Adobe, Greenfield says pulling in these partnership models reflected an evolution to how creative professionals are working with AI today. He says that customers want access to a wide variety of AI models, especially as these technologies quickly advance. This is similar to the multi-modal approach most CTOs and chief information officers have embraced when deploying AI coding tools for software developers or the application of other uses of AI in marketing, legal, and communications to improve worker productivity.
Adobe has added content credentials to make it clear to marketers when the assets they are creating are safe to use for commercial production (with Firefly) versus for ideation purposes (the external partner models). Customers have the final say on what path works best for them.
“We have a lot of customers who have different opinions on when to use different types of models and how they feel about commercial safety,” says Greenfield. “A lot of them feel that in ideation, they’re open to using anything.”
John Kell
Send thoughts or suggestions to CIO Intelligence here.
Fortune Fortune AIQ . Our third collection of stories- How Walmart, Amazon, and other retail giants are using AI to reinvent the supply chain—from warehouse to checkout. Read more
- Meet the legacy players and upstarts using AI to reinvent the energy business. Read more
- AI isn’t just entering law offices—it’s challenging the entire legal playbook. Read more
- How a bulldozer, crane, and excavator rental company is using AI to save 3,000 hours per week. Read more
- AI is already touching nearly every corner of the medical field. Read more
