[License-review] ModelGo Attribution License, Version 2.0 FINAL CALL FOR COMMENTS

[McCoy Smith]

All:

There are two ModelGo licenses currently up for consideration:

1. ModelGo Zero (basically, a -0 version of these licenses, with no 
obligations on licensee);

2. ModelGo Attribution (basically a -BY version of these licenses, with 
certain attribution obligations on the licensee).

I'd like to have those interested in either: a) approval of AI-directed 
licenses; or b) approval of licenses in general to have one more chance 
to comment upon these. Please don't assume comments you may have made in 
2025 are applicable to these licenses since they have changed quite a 
bit since then, and a copyleft version has been dropped from 
consideration due to concerns raised at the time they were submitted.

To ease review, I've done a redline of these two licenses against one 
another, here: 
https://github.com/McCoySmith/OSI-License-Review-Compare/raw/refs/heads/main/ModelGo%20Attribution%202.0%20compare%20ModelGo%20Zero%202.0.odt

This shows how the Attribution version of the license adds additional 
requirements versus the Zero version.

A few questions I'd specifically like to direct to the 
commenting/reviewing community on this list:

1. These licenses trigger obligations (to extent there are any) upon 
Distribution, which includes "providing the Licensed Materials or any 
part thereof as a hosted service or remotely accessible service, such as 
API-based access or web access." This sort of requirement is not 
uncommon in OSI approved licenses (for example, both versions of AGPL 
and the various versions of the Open Software License). Is this sort of 
obligation problematic for licenses designed for AI-models vs software? 
Is there a Freedom Zero issue with this sort of clause for an AI model?

2. The definition of "Licensed Materials" includes both the Model as 
well as "Complimentary Materials" defined as software used with the 
model. Is combining these two types of material together into the 
definition of what is licensed problematic? Some of the comments on 
OpenMDW-1.1 seemed to indicate some were concerned with that, but I'd 
like to know if in the case of these two licenses that concern attaches.

3. The Attribution version of this license has a termination clause, 
that allows for termination "if You initiate any legal action against 
the Licensor alleging that the Licensed Materials and/or Derivative 
Materials infringe any patent worldwide." Does this create a OSD 5 
problem because it only applies to Licensor, not Licensees (or just to 
specific works rather than specific parties)? Also, is this problematic 
in that a patent assertion can terminate all rights, including copyright 
rights (keeping in mind that at least MPLv2 and GPLv3 do the same)?

4. The termination for patent assertion clause applies to patent 
assertions against "Derivative Materials" -- i.e., subsequent 
modifications made by others. Is this problematic? This was generally 
considered to be too broad at the time that various patent assertion 
termination clauses were being drafted as it allowed for termination 
beyond the software the licensor originally licenses (one of the main 
concerns with how some people were reading "Liberty or Death" in GPLv2 
circa mid '00s). Most of the OSI-approved licenses that have patent 
assertion termination clauses limit them to assertions against the 
licensed work, not subsequent derivatives.

Plus one question for the drafter, Dr. Duan:

There is a difference in the language in the preamble of ModelGo Zero vs 
ModelGo Attribution. ModelGo Zero specifically points out the Sec 3 
disclaimer, but Attribution does not. Also Attribution says that if you 
don't agree to the terms, then no rights are granted; ModelGo Zero 
doesn't require this. Is there a reason for that discrepancy between 
licenses? Should they be synchronized in this section of the license to 
be identical?

On 12/8/2025 6:51 AM, Moming Duan wrote:
> Dear OSI License Review Community,
>
>
> According to the previous discussions with Pam and McCoy, I have made 
> quick updates to the MG0 and MG-BY licenses. The major modification is 
> the exclusion of models created through distillation from the 
> definition of Derivative Materials.
>
> The newest text files are attached, and as suggested, I have included 
> word-diff comparisons with the previous versions. As I mentioned in my 
> last email, I intend to withdraw MG-BY-SA at this time because I 
> cannot identify a clear gap that this license would fill, and also due 
> to the “Preferred Form” concerns raised earlier by Carlo.
>
> Please feel free to share any comments.
>
>
> Word-diff:
> MG0: https://www.diffchecker.com/f8OumPRF/
> MG-BY: https://www.diffchecker.com/jVBWA9WF/
>
>
> Best,
> Moming
>
>
>
>
> _______________________________________________
> The opinions expressed in this email are those of the sender and not necessarily those of the Open Source Initiative. Communication from the Open Source Initiative will be sent from an opensource.org email address.
>
> License-review mailing list
> License-review at lists.opensource.org
> http://lists.opensource.org/mailman/listinfo/license-review_lists.opensource.org