Intellectual Property for Marketers: Copyright, Trademark, Fair Use
Getty Images runs 24/7 web crawlers that compare live websites against its asset database, and license mismatches trigger automated demand letters ranging from $750 to over $10,000 per image. "I found it on Google" has never been a legal defense, and in 2026 the bots checking for it never sleep.
Quick Summary
- Copyright protects images, music, and video the moment they are created. Finding something via a search engine tells you nothing about your right to use it.
- Stock licenses, royalty-free libraries, and fair use are three different legal categories, mixing them up is how demand letters happen.
- Quoting or parodying a competitor can be protected commentary, or it can be infringement, the four-factor fair use test decides which.
- The US Copyright Office ruled in January 2025 that AI-generated output alone, even from detailed prompts, is not eligible for copyright protection.
- This lesson pairs with our trademark and naming lesson, that one covers protecting your own brand name, this one covers using other people's content.
Images, Music, and Video: Licenses Are Not Interchangeable
Every photo, song, and video clip is copyrighted automatically the second it is created, no registration or copyright symbol required. What you are buying when you license one is a narrow set of permissions, not ownership.
Royalty-free does not mean free. It means you pay once and can reuse the asset across many projects, but the license still has limits on scope, like editorial-only use or a cap on print runs.
Rights-managed licenses are priced per specific use: one campaign, one region, one time window. Reuse it in a second campaign without paying again and you are now infringing.
Fair use is not a license at all. It is a legal defense you can raise if sued, not a permission slip you get in advance, and it rarely covers straightforward commercial use of someone's photo in an ad.
"I found it on Google Images" describes how you located a file, not your legal right to use it. Google Images indexes copyrighted work the same way it indexes everything else, search results carry no license.
Getty Images, now valued at nearly $1 billion in 2025 revenue and expanding further after a $3.7 billion Shutterstock acquisition cleared US antitrust review in February 2026, built an entire enforcement business on this confusion. Its automated detection has occasionally flagged images it does not even own the rights to, so keep purchase receipts and license terms on file, not just a memory of "we paid for that one."
When Referencing a Competitor Crosses Into Infringement
Marketers reference competitors constantly: comparison pages, "why we're different" posts, parody ads. Some of this is protected speech, and some of it is a lawsuit.
Courts use a four-factor fair use test to decide which side of the line a use falls on:
Parody gets real protection because it is inherently transformative, it comments on the original rather than substituting for it. But the protection is not automatic. A February 2025 Delaware ruling, Thomson Reuters v. ROSS Intelligence, found no fair use where the resulting product competed directly against the original, the fourth factor sank the defense even though other factors were debatable.
The practical test for a marketing team: are you commenting on or critiquing the competitor's product, or are you just repackaging their content to save yourselves the work? The first survives scrutiny far more often than the second.
Before publishing a competitor-referencing campaign, ask whether it adds new meaning or commentary, or whether it just uses their asset because making an original one is more work. That question predicts the legal risk better than any checklist.
AI-Generated Content: Copyright Status Is Still Unsettled
Marketing teams now generate images, copy, and video with AI tools daily, and the copyright status of that output is genuinely murky, not fully resolved.
The US Copyright Office's Part 2 report, released January 29, 2025, drew one clear line: purely AI-generated output, even from detailed and carefully engineered prompts, does not qualify for copyright protection. Prompts alone are not considered enough human creative control.
The report also drew a second line that matters more for day-to-day work: if a human meaningfully selects, arranges, or modifies AI output, that human contribution can be protected on a case-by-case basis. A raw AI image is unprotected; the same image after significant human editing, curation, and arrangement may not be.
This has a direct business consequence: content your team cannot copyright is also content a competitor can legally copy. If a hero campaign asset is 100% unedited AI output, you may have no legal grounds to stop someone else from reusing it. The Copyright Office has said a Part 3 report on AI training-data licensing is still coming, so expect this area to keep shifting.
This lesson is educational, not legal advice. Copyright and fair use law turns on specific facts, jurisdiction, and how a work was actually created and modified, only a licensed IP attorney can assess your specific asset, campaign, or dispute. Treat this as background before that conversation, not a substitute for it.
Key Takeaways
- Copyright exists automatically on creation. A search engine result is not a license, regardless of how easy the image was to find.
- Royalty-free, rights-managed, and fair use are three distinct legal categories, know which one actually covers your use before you publish.
- The four-factor fair use test, purpose, nature, amount, market harm, decides whether referencing a competitor is commentary or infringement. Market harm is often the deciding factor.
- AI-generated output alone is not copyrightable as of the Copyright Office's January 2025 ruling; meaningful human editing and arrangement is what can earn protection.
- For protecting your own brand name specifically, see the companion lesson on trademark and brand protection basics, this lesson covers using other people's IP, that one covers securing yours.