Growing responsibly in the age of AI: Adobe Firefly and Stock

Image source: Adobe Stock/Robert Kneschke.

It’s been over a year since generative AI began its path into mainstream conversations and creative workflows, and during this time we have witnessed rapid increases in technical capabilities and mainstream adoption across the world. The future role of generative AI across creative segments, and how this technology will impact creators’ livelihoods continue to be at the forefront of our discussions and strategic planning.

Ensuring that we approach generative AI in a thoughtful, transparent, and responsible way is critical to our mission of connecting businesses with the best content and creators in the world, in a way that helps to empower creators with more and better ways to earn a living with their work.

Adobe Firefly: Designed to be commercially safe

Adobe Firefly, our family of creative generative AI models, was trained on licensed content, such as Adobe Stock. We designed Firefly to generate content for commercial use that does not infringe on copyright and other intellectual property (IP) right such as trademarks and logos. Customers can use Firefly-generated content with confidence for commercial purposes because we utilize a multi-layered, continuous review and moderation approach to block and remove content that violates Adobe’s policies. Adobe also offers customers IP indemnification for Firefly generated content.

Growing responsibly

In the last 12 months, Adobe Stock submissions, licenses, and contributor payouts have hit an all-time high while new contributors are signing up and generating earnings at an accelerated rate. The combination of generative AI and Stock is enabling our customers to accelerate ideation, discussion, and decision making. Ultimately, Adobe Stock customers seek the perfect image for their project that balance composition, relevant trends, commercial viability, and impact — and these are areas the stock contributor community understands deeply.

As we watch the Stock business grow and the adoption of generative AI increase, we understand the importance of helping to ensure we continue to create trust and transparency around generative AI. Over the last several months we have made significant changes to our site, policies, and procedures. Some of those changes include:

• Enhancing our submissions guidelines:
  • Contributors must not submit generative AI content with titles that imply that it is depicting an actual newsworthy event
  • Contributors must not submit work created with prompts referring to people, places, or property unless they have the legal right to do so (e.g., via a model or property release)
  • Contributors must acknowledge (via a check box) that content was "Created using generative AI tools” and that the "People and Property are fictional" if the image or video features a fake person or fake property before the content is made publicly available

• Scaling moderation capabilities by:
  • Doubling the size of our expert human moderation team who specialize in content review
  • Creating and training the team on new generative AI specific policies and best practices
  • Introducing specific generative AI moderation systems to efficiently categorize and review content for any potential policy violations
  • Refining our audit systems to monitor the existing collection more efficiently for content that may violate our new generative AI policies
  • Reviewing and removing thousands of existing assets from Stock and continuing to actively monitor the full Adobe Stock collection

• Updating our usage policies   to make it clear that:
  • Editorial uses of Adobe Stock content must not be used to mislead or deceive as doing so is a violation of our terms of service
  • Stock content should always be clearly marked when used in editorial content to help ensure people are not misled into thinking a real event is being depicted by the stock content
  • If a publisher or other licensee is found to be using a clearly-labeled Stock asset in misleading ways to deceive people, we will review and take appropriate action including terminating their rights to that Stock asset

• More clearly labeling images created with generative technology by:
  • Having a generative AI specific watermark on all generative AI content
  • Clearly indicating an image was created using generative AI
  • Enabling search filters that allow users to identify generative AI content

These changes reflect our team’s concerted effort to maintain the integrity of our offerings while addressing the evolving landscape of digital content creation and consumption as generative AI content continues to transform the way we work, create, and communicate.

Putting creators first

In addition to continuing to build a robust marketplace of content that meets our customers’ needs, we continue to explore ways in which our Stock contributors and the broader creator community can increase their earnings potential through Adobe.

We believe that generative AI technology should support creators and allow them to monetize their talents. Adobe’s approach to generative AI is built on more than a decade’s experience integrating AI into our products. By embedding Adobe Firefly natively into our products, creators can reduce and even eliminate busy work, expand their ability to create across more media, techniques and surfaces, and have more time to ideate and deliver better work.

Through the Adobe Stock marketplace, we compensate Stock contributors through paid programs like Stock royalties, Firefly contributor bonus, and payments for custom content. We also believe the law should protect creators’ styles. The new Federal Anti-Impersonation Right (FAIR Act) we’ve proposed would address intentional style impersonation for commercial gain using AI tools and would allow artists to take action against those who are intentionally and commercially impersonating their work or likeness through AI tools.

Efforts such as this include leveraging Firefly to potentially lower production costs and increase submissions, understanding demand trends with more clarity, and increasing their earnings potential through initiatives like Adobe Stock Missions (targeted “asks for content” that yield direct earnings) and freelance hiring through Behance.

At Adobe, we are committed to advancing AI in alignment with our AI Ethics principles of accountability, responsibility and transparency. Responsibility and safety are not an afterthought but are deeply integrated at every step of our product development and partnering process. We are working to create a future where creators' rights are protected, there is transparency around how content is made and edited, and customers have the choice to use gen AI models that are designed to be commercially safe as well as the option to leverage third-party models safely in our tools and services.