US Copyright Office Releases AI-Art Guidance

The US Copyright Office released a new report that deems works made using AI tools copyrightable only if they’re used to enhance the creative process of actual humans.

The Big (Generated) Picture: Copyright in the age of AI has been a hotly debated issue over the past few years. The Copyright Office’s call for public input received 10,000 comments, mostly from artists concerned about the impact on their livelihoods. The new report (the second of a planned three) gives every creative industry the greenlight on how to utilize the tech moving forward.

Behind the Prompts: The Copyright Office is taking a “remix” approach to granting copyright for AI works.

This includes:

• Artists who use AI to augment their creative work, including “selecting” and “arranging” AI-generated elements, because the final outcome is the vision of human authorship and expression.
• That also includes using tools to edit works, such as employing AI to refine actors’ accents in The Brutalist, de-aging actors in Here, using AI to optimize post-production workflows, and significantly altering AI-generated images.
• But, just entering text prompts to generate an artwork is not copyrightable, because the output is simply “the user’s acceptance of the AI system’s interpretation, rather than authorship of the expression it contains.”

The Future: Though the Copyright Office’s guidance favors human creators, it’ll likely still spark debate among creatives, many of whom view any use of AI as a faux pas. At the same time, it may push those same creatives to see how they can optimize their work — not offload their creativity — using the tech. But with the Office’s upcoming third report to center on the legality of using AI models that have been trained on copyrighted work without a license, the future of using AI may be in using systems that have been only trained on works you’ve made or have ownership of.

That would be the best of both worlds.