If you use AI-generated code, you currently cannot claim copyright on it in the US.
-
FWIW I'm not a lawyer and I'm not recommending that you do this.
Even if companies have no legal standing on copyright, their legal team will try it. It *will* cost you money.But man, oh man, I'm gonna have popcorn ready for when someone inevitably pulls this move.
@jamie also "Public Domain" doesn't exist in many juristictioms and in places like #Germanybit's the opposite and non-copyright-able code would've to be evidenced as such.
-
R relay@relay.infosec.exchange shared this topic
-
FWIW I'm not a lawyer and I'm not recommending that you do this.
Even if companies have no legal standing on copyright, their legal team will try it. It *will* cost you money.But man, oh man, I'm gonna have popcorn ready for when someone inevitably pulls this move.
The post takes real legal uncertainty and cranks it up to maximum alarm. The actual state of things is more nuanced and still evolving — courts are working through these questions case by case, and the Copyright Office indicated that a future may arise in which prompts could "sufficiently control expressive elements in AI-generated outputs to reflect human authorship."
So it's worth being thoughtful about this stuff, but "your vibe-coded repo is public domain" is not an accurate summary of the law.
-
@jamie also "Public Domain" doesn't exist in many juristictioms and in places like #Germanybit's the opposite and non-copyright-able code would've to be evidenced as such.
@kkarhan Yeah, this is very US-focused. I haven't worked with any lawyers outside the US and I'm not familiar with how copyright works outside the US at all.
However, if the company is in the US and they don't have a huge international presence, they probably aren't able to take legal action anyway.
-
The post takes real legal uncertainty and cranks it up to maximum alarm. The actual state of things is more nuanced and still evolving — courts are working through these questions case by case, and the Copyright Office indicated that a future may arise in which prompts could "sufficiently control expressive elements in AI-generated outputs to reflect human authorship."
So it's worth being thoughtful about this stuff, but "your vibe-coded repo is public domain" is not an accurate summary of the law.
For more information:
Can AI-Generated Work Be Copyrighted? New Guidance from the Copyright Office Provides Clarity
In January 2025, the Copyright Office released Part 2 of its regulatory guidelines on copyright protection and artificial intelligence (AI).
(www.mcbrayerfirm.com)
-
The post takes real legal uncertainty and cranks it up to maximum alarm. The actual state of things is more nuanced and still evolving — courts are working through these questions case by case, and the Copyright Office indicated that a future may arise in which prompts could "sufficiently control expressive elements in AI-generated outputs to reflect human authorship."
So it's worth being thoughtful about this stuff, but "your vibe-coded repo is public domain" is not an accurate summary of the law.
@tuban_muzuru There's no alarm in what I said.
-
@tuban_muzuru There's no alarm in what I said.
You're attempting to say " If you fail to disclose/disclaim exactly which parts were not written by a human, you forfeit your copyright claim on *the entire codebase*
I'll be gracious and say that's not what the law says, and if you want, I can be a jackass about this because it's not true and the last thing this place needs is yet another Chicken Little making absurd claims.
-
For more information:
Can AI-Generated Work Be Copyrighted? New Guidance from the Copyright Office Provides Clarity
In January 2025, the Copyright Office released Part 2 of its regulatory guidelines on copyright protection and artificial intelligence (AI).
(www.mcbrayerfirm.com)
@tuban_muzuru The link I posted was updated more recently than this post.
I'm also fairly certain that sufficient scrutiny will show that you cannot control the LLM to the degree that could justify copyright claim.
-
You're attempting to say " If you fail to disclose/disclaim exactly which parts were not written by a human, you forfeit your copyright claim on *the entire codebase*
I'll be gracious and say that's not what the law says, and if you want, I can be a jackass about this because it's not true and the last thing this place needs is yet another Chicken Little making absurd claims.
@tuban_muzuru You find that alarming? I meant it like "Holy shit, this is an interesting thing I hadn't considered."
And it's not an absurd claim. It's pretty much exactly what someone will say in court.
-
@tuban_muzuru You find that alarming? I meant it like "Holy shit, this is an interesting thing I hadn't considered."
And it's not an absurd claim. It's pretty much exactly what someone will say in court.
Holy shit, dude, get a subscription to Claude and be thereby instructed. If you're a competent programmer, you will be asking the LLM for spec. If you're a dumbass, you'll be asking for implementation.
-
If you use AI-generated code, you currently cannot claim copyright on it in the US. If you fail to disclose/disclaim exactly which parts were not written by a human, you forfeit your copyright claim on *the entire codebase*.
This means copyright notices and even licenses folks are putting on their vibe-coded GitHub repos are unenforceable. The AI-generated code, and possibly the whole project, becomes public domain.
Source: https://www.congress.gov/crs_external_products/LSB/PDF/LSB10922/LSB10922.8.pdf
Big bidness lobbyists will be descending on Capitol Hill to fix that.
-
Holy shit, dude, get a subscription to Claude and be thereby instructed. If you're a competent programmer, you will be asking the LLM for spec. If you're a dumbass, you'll be asking for implementation.
@tuban_muzuru My company pays for both Claude and Cursor. I use both and my team discusses these tools daily.
I'm also very familiar with how LLMs work down to a mechanical level. I've trained models from scratch with code I wrote by hand and used those models to do some pretty neat things.
I assure you, you don't have a leg up on me in this discussion.
-
Big bidness lobbyists will be descending on Capitol Hill to fix that.
@huntingdon I have no doubt they're already there. They probably don't even need to walk through the front door.
-
@tuban_muzuru My company pays for both Claude and Cursor. I use both and my team discusses these tools daily.
I'm also very familiar with how LLMs work down to a mechanical level. I've trained models from scratch with code I wrote by hand and used those models to do some pretty neat things.
I assure you, you don't have a leg up on me in this discussion.
I can assure you, Jamie, whatever talents you may possess, possession of a Juris Doctor is not one of them. So do us all a favor and quit handing out what I believe to be so much Chicken Little-ing.
-
I can assure you, Jamie, whatever talents you may possess, possession of a Juris Doctor is not one of them. So do us all a favor and quit handing out what I believe to be so much Chicken Little-ing.
@tuban_muzuru I don't have a JD, that's correct. I do, however, have two siblings that do and this entire thread came out of a conversation with one of them.
And no, I will do you no favors. You came into *my* mentions. I had no idea you existed 40 minutes ago. So I owe you nothing. You'll just have to find a way to accept that I will continue posting my thoughts on my own account on my own server.
-
Holy shit, dude, get a subscription to Claude and be thereby instructed. If you're a competent programmer, you will be asking the LLM for spec. If you're a dumbass, you'll be asking for implementation.
@tuban_muzuru @jamie if youre a competent programmer you will absolutely not be asking an LLM for spec. you will read documentation.
-
@tuban_muzuru @jamie if youre a competent programmer you will absolutely not be asking an LLM for spec. you will read documentation.
@oay @tuban_muzuru This is the way.
-
@tuban_muzuru I don't have a JD, that's correct. I do, however, have two siblings that do and this entire thread came out of a conversation with one of them.
And no, I will do you no favors. You came into *my* mentions. I had no idea you existed 40 minutes ago. So I owe you nothing. You'll just have to find a way to accept that I will continue posting my thoughts on my own account on my own server.
Shrug. Here's a tip - when you put up a para like this one: "It'll be interesting to see what happens when a company pisses off an employee to the point where that person creates a public repo containing all the company's AI-generated code. I guarantee what's AI-generated and what's human-written isn't called out anywhere in the code, meaning the entire codebase becomes public domain."
- I can make the observation you're being a Chicken Little. You guaranteed it.
-
@tuban_muzuru @jamie if youre a competent programmer you will absolutely not be asking an LLM for spec. you will read documentation.
-
@tuban_muzuru @jamie not with LLM
-
Shrug. Here's a tip - when you put up a para like this one: "It'll be interesting to see what happens when a company pisses off an employee to the point where that person creates a public repo containing all the company's AI-generated code. I guarantee what's AI-generated and what's human-written isn't called out anywhere in the code, meaning the entire codebase becomes public domain."
- I can make the observation you're being a Chicken Little. You guaranteed it.
@tuban_muzuru The guarantee is that the parts of the codebase that aren't written by humans are not called out. You can read into the rest however you like.
