A few notes to start with. The U.S. Copyright Office and The New York Times are having what amounts to a legal slap fight about AI-generated content. The Copyright Office says AI is too random to copyright, while the NY Times is suing OpenAI because their AI is copying articles too perfectly. They can’t both be right… or can they?
Why is this important?
- The Copyright Office just dropped their 2025 report saying AI outputs are like rolling dice – totally random and unpredictable
- The NYT lawsuit claims OpenAI’s models can spit out their articles word-for-word
- If you’re creating content using AI (like I do), this affects whether you can copyright your work
The U.S. Copyright Office View – Rolling the Dice The Copyright Office’s stance is pretty straightforward: AI-generated works can’t be copyrighted because they lack human authorship. They compare it to rolling dice – you might influence how the dice are thrown, but you can’t control exactly how they land.
The NYT View – It’s Not Random If You Can Copy It Here’s where it gets interesting. The NYT is suing OpenAI because their AI can reproduce NYT articles verbatim or nearly verbatim. That’s about as far from random as you can get. It’s like claiming your dice always roll snake eyes when you want them to.
A Different Perspective – The Photoshop Analogy Here’s something to think about: When an artist uses Photoshop brushes (which often have randomized elements), they can still copyright their work. Why? Because they’re guiding the process. They’re using the tool to express their vision.
If I use AI to create content, and I can guide it to express my vision through careful prompting, isn’t that the same thing? I’m not just throwing prompts at the wall and seeing what sticks – I’m crafting and refining until I get exactly what I want.
Let’s break down the two conflicting viewpoints:
| Viewpoint 1: AI is Unpredictable (Copyright Office) |
Viewpoint 2: AI is Controllable (NYT Lawsuit) |
|---|---|
| AI models function like random generators, so their outputs cannot be copyrighted. | AI is capable of producing controlled, replicable outputs, meaning it behaves like a tool rather than an independent creator. |
| Prompts are not enough to establish authorship because they lack sufficient control over the outcome. | If AI can be steered to produce specific outputs (even replicating copyrighted content), then it is not purely random—and the prompter’s role should be recognized. |
| AI content is too unpredictable to assign ownership. | If OpenAI is responsible for infringement, then AI must be predictable enough to claim authorship. |
The Twist If the NYT wins their lawsuit by proving AI can predictably reproduce content, they might accidentally help creators copyright their AI-generated work. After all, if AI isn’t just “rolling the dice,” then maybe it’s more like a Photoshop brush – a tool that helps express the creator’s vision.
What This Means For Content Creators If you’re using AI to create content (like many of us are), keep an eye on this case. A NYT victory could mean:
- The Copyright Office might have to revise their “it’s all random” stance
- AI-assisted works might become eligible for copyright protection
- Your carefully crafted prompts might actually count as authorship
The Bottom Line We’re in uncharted territory here. The old rules about copyright were written for a world where content creation was purely human. Now we’re trying to apply those rules to AI-assisted creation, and it’s about as awkward as trying to explain TikTok to your grandparents.
One thing’s for sure – if the NYT proves AI isn’t random, they might just help every AI content creator claim copyright on their work. Sometimes the law of unintended consequences works in our favor.
What do you think? Should AI-assisted works be copyrightable if the creator exercises meaningful control? Let me know in the comments below.
(Note: I use AI in my content creation process, and yes, I see the irony in using AI to write about AI copyright. It’s like inception, but with more legal paperwork.)
