We aren't even through January and there has been a flurry of genAI launches and surprises and disruptions. Just as marketers are finding their feet to scale more into this space in adopting genAI tools and workflows, one nagging concern I that they (and we) all share is brand safety. Not just making sure that your ad doesn't show up in the wrong place - I mean that when your content arsenal expands by 10x, or more likely by 1000x there are new risks: How are you protecting the works of your brand, how are you ensuring you are not infringing the rights of others, how are you protecting and growing your own IP?

Well, enter the next January talking point: a 41-page U.S. Copyright Office report that's about to rewrite the rules of creative production. Released just yesterday, the Copyright Office's "Copyright and Artificial Intelligence, Part 2: Copyrightability" report is a manifesto for the future of creativity in the age of generative AI.

Let's break down the report's core findings - and what they mean for your brand's creative strategy:

The Human Authorship Principle

The report makes one thing brutally clear: copyright protection requires human creativity. This isn't a suggestion - it's a fundamental legal principle. Those AI-generated images, texts, and designs you've been collecting? They're essentially creative placeholder cards - interesting, but legally valueless.

The report cites a landmark case from the District of Columbia court, which stated unequivocally that "copyright law protects only works of human creation" and that "human authorship is a bedrock requirement of copyright."

The Prompt Paradox

Here's where it gets interesting. Many marketers have been banking on sophisticated prompt engineering as a path to copyright protection. The Copyright Office has a definitive response: Not happening.

Even the most elaborate prompts are considered "unprotectable ideas." As the report bluntly states, prompts "essentially function as instructions that convey unprotectible ideas." Whether your prompt is a one-word request or a 500-word detailed description, the result remains the same: the output is not yours.

Here's a twist many brands haven't considered: Just because AI-generated content isn't copyrightable doesn't mean it's legally safe to use.

The Copyright Office report reveals a crucial nuance: While prompt-based AI outputs aren't protected by copyright, they could still potentially infringe on existing copyrights. Why? Because AI models are trained on massive datasets of existing creative works.

This means that seemingly "original" AI-generated content might actually be:

  • Algorithmically reconstructing elements from copyrighted works

  • Inadvertently reproducing protected visual or textual elements

  • Creating derivative works without proper licensing

Practical Risks for Brands:

Using AI-generated content could expose you to unexpected legal challenges and your "original" AI image might contain traces of copyrighted material. Prompt-based generation doesn't guarantee originality or safety. Net/net: Your AI-generated content isn't just uncreative - it could be a legal minefield.

Here is where I must give a shout out to Adobe. I spent a lot of time with the Adobe team at CES. They have faced criticism that they have been slower to market than others with the GenAI suite and the developments of Firefly in particular. But this is why: while there are 'wild west' entrants saying they will underwrite the legal risks from infringing copyrighted material in your genAI work, that's hardly the level of brand safety standards that we need to all hold ourselves to. Adobe focuses on copyright and IP protection of your works, and the works of others, and properly documents the human role in creativity. Better to be a little late, and safe.

The Tools of Transformation

But there is a gift for us here - some guardrails that have been needed. The report outlines clear pathways for copyright protection:

1. AI as a Creative Assistant: When AI functions as a tool that helps realize human creativity, copyright protection remains intact.

2. Substantial Human Modification: If you significantly transform AI-generated content, adding your unique creative expression, you can claim copyright on those modifications.

3. Expressive Inputs: When you input your own copyrightable work into an AI system, and that work remains perceptible in the output, you maintain copyright protection.

The Global Context

The report doesn't just speak to U.S. law - it provides a fascinating global perspective. Other countries are wrestling with similar challenges:

  • Korea allows copyright only when a human contributes creatively to the expressive form.

  • Japan evaluates copyrightability based on human input, generation attempts, and subsequent modifications.

  • European Union members largely agree that current copyright principles are sufficient, with protection granted only when human input is "significant."

Practical Playbook: Protecting Your Brand's Creative Assets

Let's get brutally practical. Here's exactly what brand owners and agencies need to do to maintain legal protection and creative integrity when using generative AI:

1. Documentation is Your Defense

Create a detailed 'creative process log' for every AI-assisted project and then record exactly how AI outputs were modified. Ensure that your workflows capture original AI generation and subsequent human transformations to identify specific human creative decisions and modifications.

2. Workflow Transformation

Establish a mandatory 'human intervention' stage for all AI-generated content and overall, require at least 30-40% substantive creative modification. Train teams to see AI as a starting point, not a finish line. To be clear - this is not about significant human work on iterations of a work, but the main idea that these iterations are then dynamically rendered from.

3. Creative Modification Guidelines

Go beyond surface-level edits to demonstrate meaningful creative choices:

  • Restructuring content

  • Rewriting substantial portions

  • Adding unique conceptual layers

  • Introducing brand-specific storytelling elements

Develop a formal AI content review process that includes creating templates for documenting AI tool usage, establishing clear guidelines for acceptable AI assistance. Stay in close consultation with legal teams to develop specific brand protocols.

So, Is This Good or Bad News?

For progressive brands (and our clients ) this is excellent news. It reinforces the value of human creativity, it creates a clear competitive advantage for agencies and brands willing to truly innovate and it prevents commoditisation of creative work by encouraging deeper, more meaningful use of AI tools.

For lazy marketers, this is a wake-up call. AI is not a shortcut, creativity cannot be completely automated and your unique human perspective is your most valuable asset.

Overall, this isn't about restricting AI. It's about elevating how we use it. The Copyright Office has essentially said: "Prove your creativity. Show your work. Make something uniquely yours." Pretty fair, actually.

Your AI workflow should look less like 'generate and go' and more like 'collaborate, transform, elevate'. Creativity is not about generating content. It's about meaningful human intervention, about taking raw material and transforming it through unique vision, context, and intention.

AI hasn't replaced the creative spark - it's become a new kind of creative kindling. Your job isn't to generate content. Your job is to create fire.

Have a think about incorporating the three pillars of AI-Enabled Creativity

1. Provocation: Challenge AI systems with complex, nuanced inputs

2. Transformation: Radically reshape AI-generated content

3. Orchestration: Create systematic approaches to AI-human collaboration

As the report concludes, existing legal principles can adapt to technological change. Your creative strategy should do the same.

Brandflow is written by Justin Billingsley, who has spent his career on all three sides of the industry's table: senior client, global agency leader, technology founder. First published 30 January 2025 in the Brandflow newsletter on LinkedIn.