[License-discuss] [SUBMISSION] AI-MIT License 1.0 — permissive license for AI-generated code

Tue Mar 17 19:38:18 UTC 2026

This is why I started deprecating Open Hardware licenses. Schematics are
not currently copyrightable, and what if they were? Perhaps Horowitz and
Hill's "Art of Electronics" textbook would be collecting royalties from
every electronic product worldwide.

Bruce Perens K6BP

On Tue, Mar 17, 2026, 11:25 Henrik Ingo <henrik.ingo at gmail.com> wrote:

> Hi all, it's been a few years since I was active here, have been heads
> down in startup mode.
>
> I have a policy opinion, which is independent of this particular license
> submission,but very relevant:
>
> OSI should not approve any licenses that are specifically targeting code
> produced by an AI, and should be careful in adding clauses to existing OSS
> licenses as well. (The latter may turn out to be necessary, but if it
> becomes necessary, OSI should first go through clear policy guidelines on
> wording that may and may not be used in such clauses.)
>
> As Nik also correctly observes, code generated by an AI is in general
> considered to not be copyrighted at all. (This is a) because legislators
> were of the opinion a machine or software program cannot, by definition,
> produce a creative work , and b) even if they could, they are not human so
> they cannot own property, material or immaterial. For the same reason
> animals cannot own copyright either.)
>
> Projecting into a very near future, most code will be generated by AI, and
> therefore most code will not be copyrighted. If we go back to our roots of
> our movement, with Stallman and a printer driver and so on... this is the
> dream come true! We should embrace and defend this future, and I for one am
> thankful it is happening in my lifetime.
>
> Approving and using a license such as the one proposed in this thread
> would be a terrible own goal to make at this point. OSS licenses mostly
> deal with copyright, so since there is no copyright, there is no need for a
> license. In particular, using words like "permission is hereby granted" is
> hugely misleading: The user of the software does not need to ask for
> permission, and in any case the person putting the license into the git
> repository is not in a position to grant such permission. However, doing so
> could very well create precedent, or a habit, that some kind of license is
> still needed, and that could be misused by a proprietary software company
> to claim that their restrictive license must also be followed.
>
> Finally, a good reason for the OSI not to approve such licenses - at least
> until further discussion - is that it is out of scope. OSI does not have a
> mandate to review or approve such licenses. Since AI generated code is not
> copyrighted, it isn't really open source nor closed source. (It might be
> the latter, but certtainly isn't open source.) Since it isn't open source,
> it is not within OSIs mandate.
>
>
> I agree that it might become necessary to make some clarifying statement
> about the above situation. As such, OSI may want to recommend some short
> informational statement explaining that some body of code was fully or
> mostly generated by AI, and therefore is not copyrighted, and therefore the
> reader is free to copy, use, etc... the program(s). And if there is such a
> statement, it could also come with a disclaimer that further emphasises the
> point by saying that since the software was not produced and is not
> copyrighted by a human, there is also no person that could be held liable
> for any kind of malfunction or damage. (Except of course if a distributor
> of the software specifically wants to offer such warranty, perhaps against
> a fee.)
>
> henrik
>
>
>
> On Sat, Mar 14, 2026 at 2:32 PM Nik <nik.sharky at gmail.com> wrote:
>
>> Apologize, my mistake. I'm new to mailing lists and sent this to both
>> addresses.
>> This proposal is currently up for discussion, and if the community
>> agrees, the final text can be submitted for review and registration after
>> discussion, as I understand it.
>>
>> Nik
>>
>> пт, 13 мар. 2026 г. в 18:32, Pamela Chestek <
>> pamela.chestek at opensource.org>:
>>
>>> McCoy, I also have an email where it was submitted to license-review on
>>> 3/12/26 at 5:30 pm PT?
>>>
>>> Pam
>>>
>>> On 3/13/2026 9:10 AM, McCoy Smith wrote:
>>>
>>> Nik:
>>>
>>> Are you requesting approval of this license by OSI, or just discussion
>>> of the license? You've submitted to the license-discuss list, which is
>>> where licenses are discussed, but does not result in the license being put
>>> through the approval process. Your statement that you are "submitting" the
>>> license "for consideration" is ambiguous.
>>>
>>> If you are seeking approval, you need to use the correct mailing list
>>> and provide all the data required for a submission, which has not been done
>>> in your e-mail below.
>>> On 3/12/2026 4:20 AM, Nik wrote:
>>>
>>> Dear OSI License Review Committee,
>>>
>>> I am submitting the **AI-MIT License, Version 1.0** for consideration by
>>> the Open Source Initiative.
>>>
>>> ## Summary
>>>
>>> The AI-MIT License is a permissive open-source license designed to
>>> address a genuine gap: existing licenses were written for human authors and
>>> handle AI-generated code poorly, creating false implications about
>>> authorship and copyright status.
>>>
>>> The license is deliberately minimal — it preserves the structure and
>>> permissiveness of the MIT License while adding three targeted changes for
>>> the AI context.
>>>
>>> ## The problem it solves
>>>
>>> 1. **False authorship implication.** When `Copyright (c) [year]
>>> [author]` is applied to fully AI-generated code, it implies human
>>> authorship and copyright that may not legally exist in most jurisdictions.
>>>
>>> 2. **No standard for disclosure.** There is no widely adopted mechanism
>>> for disclosing whether code is AI-generated, AI-assisted, or
>>> human-authored. This matters for supply-chain security, regulatory
>>> compliance (EU AI Act), and intellectual honesty in open source.
>>>
>>> 3. **Undefined copyright status.** Fully autonomous AI-generated code
>>> (no human creative input) is in a legal grey zone in most jurisdictions. A
>>> license that claims copyright over it is at best misleading, at worst
>>> invalid.
>>>
>>> ## What the license does differently from MIT
>>>
>>> The license adds one structural element (the Authorship Declaration) and
>>> three conditions/clauses:
>>>
>>> **Authorship Declaration** — a required checkbox at the top of the
>>> LICENSE file with three modes:
>>> - *Fully AI-generated*: no copyright claimed; code dedicated to public
>>> domain
>>> - *AI-assisted*: human-directed, AI-generated; standard copyright applies
>>> - *Human-authored*: AI used as a tool only; identical to MIT posture
>>>
>>> **Condition 2 — Transparency**: redistribution or use as AI training
>>> data must not misrepresent AI origin as human authorship.
>>>
>>> **Condition 3 — No Copyright Claim**: for fully autonomous code,
>>> explicit public domain dedication (with a perpetual irrevocable fallback
>>> for jurisdictions where public domain dedication is impossible).
>>>
>>> **Extended disclaimer**: adds three AI-specific disclaimers about
>>> training data provenance, regulatory compliance, and jurisdictional
>>> limitations of the authorship declaration.
>>>
>>> ## OSD compliance analysis
>>>
>>> 1. **Free Redistribution** ✓ — no restriction on sale or distribution
>>> 2. **Source Code** ✓ — no source restriction
>>> 3. **Derived Works** ✓ — modification and redistribution permitted
>>> 4. **Integrity of the Author's Source Code** ✓ — no patch-file
>>> requirement; attribution preserved
>>> 5. **No Discrimination Against Persons or Groups** ✓
>>> 6. **No Discrimination Against Fields of Endeavor** ✓
>>> 7. **Distribution of License** ✓ — same rights apply to all recipients
>>> 8. **License Must Not Be Specific to a Product** ✓
>>> 9. **License Must Not Restrict Other Software** ✓
>>> 10. **License Must Be Technology-Neutral** ✓
>>>
>>> The Transparency condition (Condition 2) requires disclosure of AI
>>> origin but does not restrict use in any field — it is an
>>> attribution/honesty requirement, not a field-of-endeavor restriction.
>>>
>>> ## SPDX identifier
>>>
>>> We are concurrently requesting the SPDX identifier `AI-MIT-1.0` through
>>> the SPDX GitHub repository.
>>>
>>> ## Repository
>>>
>>> The full license text, README, translations, and supporting materials
>>> are available at:
>>> https://github.com/ai-mit-license/ai-mit-license
>>>
>>> ## A note on meta-context
>>>
>>> This license was initially drafted with AI assistance (Claude,
>>> Anthropic) at the direction of a human. We believe this is appropriate and
>>> have disclosed it in the repository. The license is itself an example of
>>> the category of work it governs.
>>>
>>> We welcome feedback from the committee and the community at large.
>>>
>>> Respectfully,
>>> Nik
>>>
>>> _______________________________________________
>>> The opinions expressed in this email are those of the sender and not necessarily those of the Open Source Initiative. Official statements by the Open Source Initiative will be sent from an opensource.org email address.
>>>
>>> License-discuss mailing listLicense-discuss at lists.opensource.orghttp://lists.opensource.org/mailman/listinfo/license-discuss_lists.opensource.org
>>>
>>>
>>> _______________________________________________
>>> The opinions expressed in this email are those of the sender and not necessarily those of the Open Source Initiative. Official statements by the Open Source Initiative will be sent from an opensource.org email address.
>>>
>>> License-discuss mailing listLicense-discuss at lists.opensource.orghttp://lists.opensource.org/mailman/listinfo/license-discuss_lists.opensource.org
>>>
>>>
>>> _______________________________________________
>>> The opinions expressed in this email are those of the sender and not
>>> necessarily those of the Open Source Initiative. Official statements by the
>>> Open Source Initiative will be sent from an opensource.org email
>>> address.
>>>
>>> License-discuss mailing list
>>> License-discuss at lists.opensource.org
>>>
>>> http://lists.opensource.org/mailman/listinfo/license-discuss_lists.opensource.org
>>>
>> _______________________________________________
>> The opinions expressed in this email are those of the sender and not
>> necessarily those of the Open Source Initiative. Official statements by the
>> Open Source Initiative will be sent from an opensource.org email address.
>>
>> License-discuss mailing list
>> License-discuss at lists.opensource.org
>>
>> http://lists.opensource.org/mailman/listinfo/license-discuss_lists.opensource.org
>>
> _______________________________________________
> The opinions expressed in this email are those of the sender and not
> necessarily those of the Open Source Initiative. Official statements by the
> Open Source Initiative will be sent from an opensource.org email address.
>
> License-discuss mailing list
> License-discuss at lists.opensource.org
>
> http://lists.opensource.org/mailman/listinfo/license-discuss_lists.opensource.org
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensource.org/pipermail/license-discuss_lists.opensource.org/attachments/20260317/9f1f749d/attachment-0001.htm>