<div dir="ltr">Thank you for the substantive feedback. Both points are well taken and I want to<br>address them directly. I attach new license text according to raised questions and propositions.<br><br>\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501<br><br>POINT 1 \u2014 THE NAME<br><br>You are correct that "MIT" is a registered trademark of the Massachusetts Institute<br>of Technology, and using it in a new license name creates ambiguity at best and<br>trademark risk at worst.<br><br>It is good idea to rename it to:<br><br> AI-Attribution License (AIAL)<br><br>The new name reflects what the license actually does \u2014 extends attribution<br>requirements to cover AI authorship \u2014 without borrowing the MIT name. <br>It is also more descriptive for developers choosing a license.<br><br>The SPDX identifier would become: AIAL-1.0<br><br>I welcome any alternative name suggestions from the community if there are<br>objections to "AI-Attribution License" as well.<br><br>\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501<br><br>POINT 2 \u2014 PER-FILE ATTRIBUTION<br><br>This is the more substantive point and I agree with it completely.<br><br>You correctly identified that a single project-level Authorship Declaration is<br>insufficient for any real-world project with a development history. <br>Real projects can contain:<br> - files written entirely by humans (pre-AI era)<br> - files generated by AI, never touched by humans<br> - files where a human wrote 80% and AI generated 20%<br> - files inherited from third-party OSS projects<br> - files modified by multiple AI tools across multiple years<br><br>A single checkbox in a root LICENSE file cannot honestly describe this mix.<br><br>The solution, as you suggested (referencing APL2.0), is per-file attribution.<br>The good news is that the ecosystem for this already exists and is widely<br>adopted: the REUSE Specification (reuse.software, FSFE) combined with SPDX<br>per-file headers.<br><br>REUSE already defines two standard tags per file:<br><br> # SPDX-FileCopyrightText: 2026 Jane Doe<br> # SPDX-License-Identifier: GPL-3.0-or-later<br><br>The revised license extends this with <br>two new optional-but-recommended AI-specific SPDX tags:<br><br> SPDX-AI-Authorship: [fully-generated | ai-assisted | human-authored]<br> SPDX-AI-Tool: [tool name and version]<br><br>These tags are:<br> - Consistent with existing SPDX tag-value syntax<br> - Backward-compatible (they are optional for human-authored files where<br> no AI was involved at all)<br> - Machine-readable and scannable with standard grep/REUSE tooling<br> - Compatible with the REUSE specification \u2014 they sit alongside existing<br> SPDX-FileCopyrightText and SPDX-License-Identifier tags<br><br>The project-level LICENSE file retains a simplified Authorship<br>Declaration as a default/fallback for projects that have not yet adopted<br>per-file headers, but per-file headers take precedence where present.<br><br>\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501<br><br>REVISED LICENSE \u2014 AI-Attribution License (AIAL) v1.0<br><br>The revised license text is attached as AIAL-1.0.txt. Key changes:<br> 1. Name changed from "AI-MIT License" to "AI-Attribution License (AIAL)"<br> 2. SPDX identifier changed from AI-MIT-1.0 to AIAL-1.0<br> 3. Project-level Authorship Declaration retained as fallback default<br> 4. Per-file attribution via new SPDX tags defined and required for<br> files where the project-level declaration does not apply uniformly<br> 5. Precedence rule: per-file tags override project-level declaration<br> 6. Inherited OSS files: explicit provision for files where AI authorship<br> status is unknown (SPDX-AI-Authorship: unknown)<br><br>\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501<br><br>Thank you again for the careful reading. This kind of feedback is exactly<br>what makes the process valuable.<br><br>Nik<br><a href="mailto:nik.sharky@gmail.com">nik.sharky@gmail.com</a></div><br><div class="gmail_quote gmail_quote_container"><div dir="ltr" class="gmail_attr">\u043f\u0442, 13 \u043c\u0430\u0440. 2026\u202f\u0433. \u0432 18:35, Josh Berkus <<a href="mailto:josh.berkus@opensource.org">josh.berkus@opensource.org</a>>:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">On 3/12/26 5:30 PM, Nik wrote:<br>
> I am submitting the AI-MIT License, Version 1.0 for OSI approval. The <br>
> full license text is attached to this email as a plain text file (AI- <br>
> MIT-License-1.0.txt).<br>
<br>
So this is interesting, but will need some iterations I think <br>
(completely aside from any requirements the attorneys have).<br>
<br>
The first part is the name; we may not be able to call it MIT, which is <br>
after all a trademark of the folks in Cambridge and this is not their <br>
license. So we might need to call it AI-Attribution or something (which <br>
would be a good name considering the purpose of the license).<br>
<br>
The second, larger problem occurs in the actual construction of software <br>
with AI assistance. As written, this license targets only brand-new <br>
projects which are created "from scratch" and never modified again. <br>
This is fine for those, but I think project creators want to plan for <br>
success.<br>
<br>
For any substantial project which has a history of development, there is <br>
going to be a mix, including: files that are wholly AI-generated, files <br>
which are AI-assisted, files which were written by humans, and files <br>
which were inherited from other OSS projects whose licenses do not <br>
require AI attribution so we don't know. Further, files which have been <br>
modified several times are going to have been modified by several AI tools.<br>
<br>
If you want to follow this concept, I really think that some kind of <br>
per-file attribution (ala APL2.0) is going to be necessary, rather than <br>
a single statement of authorship over the whole project.<br>
<br>
-- <br>
-- Josh Berkus<br>
OSI Board Member<br>
</blockquote></div>