Here was my attempt to write a clause prohibiting language model training/inference:
https://bugfix-66.com/7a82559a13b39c7fa404320c14f47ce0c304fa...
3. Use in source or binary form for the construction or operation
of predictive software generation systems is prohibited.
The Humans Only Clause adds an explicit licensing fee, and what else?
How is the clause worded?
If that confused you, or you consider it "obnoxious", then you are not the target audience.
That's ok. Hacker News is not 100% hackers!
> The license must not restrict anyone from making use of the program in a specific field of endeavor. For example, it may not restrict the program from being used in a business, or from being used for genetic research.
It also seems to violate freedom 0 of the FSF's four essential freedoms that define free software:
> The freedom to run the program as you wish, for any purpose (freedom 0).
I'm not sure this can be used by open source projects if they want to remain open source projects.
How is this behavior different than 90% of human coders? Most SW devs scream if you ask them to pay for something, whether its apps, code, TV, movies, etc. But they will happily try to build startups on top of a vast mountain of free code. I really don't care if my code gets sucked up by the AI vacuum, humans have been doing that quite well for awhile now.
A legal grey area exists as to whether publicly available creations (code or art) can be used to train datasets for generative AI projects without infringing their creators' underlying copyrights. Other types of claims, such as violation of license agreements and DMCA violations, require proof of damages to substantiate.
The legal solution we’ve identified is to add a specific damages amount to the license itself — a licensing fee. The failure to pay such a fee would cause the creator to suffer damages in the amount of the fee. By imbedding a licensing fee into a traditional open-source license, a creator can solve the proof-of-damages issue that could otherwise limit a claim under the DMCA or for breach of contract, and limit the fee to generative AI companies.
That’s why we built the Humans Only Clause. If you don’t want your code used by Copilot in this way, the Humans Only Clause can help strengthen your protections from use for training purposes. It’s a simple addition to your existing open source license to keep it free use and open source for other developers, but to prevent use without attribution by generative AI companies.
You can access the Humans Only Clause and insert it into your GitHub repo by going to PayToTrain.ai — we also built a payments form where you can set your own licensing fee depending on how valuable you believe your repo to be. If we get enough people using this clause, there’s a good chance we can assemble a separate class for a future class action, where each user gets significantly higher damages than what’s available statutorily under existing DMCA lawsuits.
On a philosophical level, we believe that the open source community is based on principles of taking and giving back to the collective. AI-based programming assistants strip away any attribution while drawing from the underlying contributions of the community. We want the open source community to continue to be open source, but we don’t want big companies to profit on our code.
If you’re interested, check it out: paytotrain.ai. We’d love to hear what you think