New Adobe Advocates for Artists’ Control over AI Training on Creative Works, Urges Government Action

Adobe, a leading software company specializing in imaging and design, has expressed its support for granting artists the right to prevent artificial intelligence (AI) systems from being trained on their creative works. This stance was shared during a Senate Judiciary Committee subcommittee hearing on AI and intellectual property rights. Dana Rao, Adobe’s General Counsel and Chief Trust Officer, highlighted the importance of protecting creators while acknowledging the need for AI systems to be trained on extensive datasets for accurate outputs.

Adobe’s Approach to Striking a Balance between AI Training and Creative Rights: Rao informed the committee that Adobe acknowledges the significance of training AI systems on large amounts of data to ensure accurate results. However, he also stressed the necessity of implementing measures to safeguard the rights of artists and creators. In response, Adobe exclusively trained its generative AI system, Firefly, using licensed images from the company’s stock photo library and other openly licensed content. This approach guarantees that the training data is commercially secure and respects the rights of creators.

The Introduction of a “Do Not Train” Tag: Rao proposed that creators should be able to attach a “do not train” tag to their work, indicating that AI systems cannot use their creations for training purposes. He emphasized the importance of industry and government support in ensuring that AI data callers respect this tag, giving creators the option to exclude their data from AI training datasets. To reinforce this protection, Rao suggested that Congress consider making the use of such tags mandatory to safeguard artists’ works and ensure their creative rights are respected.

Preventing Unfair Impersonation: Alongside the “do not train” tag, Rao argued that Congress should establish a new federal anti-impersonation right to safeguard artists from economic harm caused by AI systems intentionally imitating their distinctive style or likeness. Granting artists legal recourse in such situations would address concerns and provide a framework for protecting their rights in the digital era. Holding individuals accountable for misusing AI tools would alleviate customer concerns and safeguard the rights of artists.

Testimony from Artist Karla Ortiz: During the subcommittee hearing, artist Karla Ortiz voiced her concerns about the impact of generative AI on artists’ livelihoods. Ortiz explained that generative AI differs from previous technologies as it exploits and utilizes the hard work, creativity, and innovation of artists without their permission, credit, or compensation. She highlighted instances where artists’ works were taken and used to train for-profit technologies with massive datasets containing billions of image and text pairs. Ortiz’s testimony emphasized the urgent need to protect artists’ rights in the face of advancing AI technologies.

Congressional Action and AI Regulation: The hearing shed light on ongoing discussions within Congress regarding AI regulation and legislation. While ideas such as creating a new federal agency or commission dedicated to handling AI have been proposed, progress in this area has been limited thus far. However, Adobe’s support for artists’ rights and its call for government action demonstrate the pressing need to address concerns related to AI-generated content and the protection of creative works in the digital age.

As the discussion around AI and its impact on intellectual property rights continues, Adobe’s support for artists’ rights and its call for government action highlight the need to address concerns related to AI-generated content and the protection of creative works. By implementing measures such as the “do not train” tag and establishing a federal anti-impersonation right, policymakers can work towards striking a balance between AI innovation and the preservation of artists’ rights in the ever-evolving digital landscape.

“Enabling Effective Regulation for AI: A Collaborative Effort”