AI tools can whip up amazing content in seconds, but they raise a thorny question: when AI is involved in creating a design or copy, who actually owns the end result? For Ghanaian and African creators, the legal landscape is murky. 

Recently the U.S. Copyright Office laid out some guidance that offers clues. They emphasize this principle: only the human creative contribution is protected. If you feed an image into an AI generator and it produces a remix, you can’t simply copyright “AI” as author. The human author must explicitly claim only what they created, which is usually the arrangement, edits or text prompts, and must disclaim any direct AI-generated parts.

That means if your agency uses an AI background generator to make a new ad visual, you should document that your copyright claim covers your own selection, coordination, and arrangement of elements, not the raw AI output. Anything above “de minimis” (significant AI-only content) must be excluded from the claim. In other words, AI isn’t an author but you are. But the flip side is: if there’s no real human creativity in there, U.S. law says there may be no copyright at all.

What about Ghana specifically? Our laws didn’t anticipate AI, but they do cover computer-generated works. The Copyright Act 2005 treats software as literary work, so it protects AI programs. For AI-generated art or text, Ghana’s law implies protection only if there’s human artistry behind it. As one analysis notes, “copyright typically applies to works created by humans” in Ghana, and if an AI merely assisted, it’s unclear—but we can adapt existing rules if the human input is significant. That mirrors the U.S. stance: the human “stitching” of AI outputs can be copyrighted, but the machine’s contribution cannot. Ghana even explicitly protects “derivative works” that show originality. So a photo or song that you edit or arrange using AI might qualify as your own creative derivative.

For brands, the takeaway is caution and clarity. If you use generative AI on your logo, advertising jingles or product copy, keep records of who did what. Disclose the use of AI and be prepared to stake copyright only in your adjustments and choices. Likewise, don’t assume AI manufacturers’ data lakes grant you rights to stuff; if an AI was trained on copyrighted African art, using its output could infringe unless you have a license. Finally, embed metadata. Tag your creations and if possible use emerging NFT-style timestamping. That way, you can prove when and how a piece was created – a smart move if disputes arise.

Raymond Klutse notes that while “AI cannot be the author,” humans always can protect the combination of AI with their own creativity. In practice, a Ghanaian designer might use AI to generate textures but then blend them with original sketches; he or she can claim copyright in the overall design because of that human touch.

The new U.S. guidance is a useful model: always specify your human contribution and exclude raw AI content. In Ghana, while the law hasn’t changed overnight, creators should proceed the same way. Use AI to speed workflows, but remember you (the team) are the creator when you add your imagination. Keep transparent notes (or contracts) about AI use. That will ensure brands and agencies can assert ownership over the final work, even in the age of the algorithm.