Does GitHub Data Grab for AI Training Violate Licenses?

July 22, 2021

Programmer Nora Tindall has taken to Twitter to call out Microsoft property GitHub on violating licenses for algorithm training purposes. She shares a screenshot of an exchange she had with GitHub Support that seems to confirm her charge:

[Tindall] I am specifically asking if any code from my GitHub account, most of which is licensed GPL, was used in the training set. It is a simple question.”

[GitHub] Sorry about the delay in getting back to you. I reached out to the team about this. Apparently all public GitHub code was used in training. We don’t distinguish by license type. I hope that answers your question!

It does indeed answer Tindall’s question, and she vows to pursue legal action. Predictably, the post prompted a flurry of comments, so navigate there to read that debate. It seems like the legality of this data usage is nebulous until courts weigh in. We note this exchange:

[Daniel Monte] Is there any precedent for training an AI on copyrighted content being a violation of said copyright?

[Nora Tindall] No, there’s no precedent in any of this. This is the deciding moment for the future of the copyleft ideal, and of free software in general. Maybe for copyright as a whole, actually, since this has applications outside software.

[Laurie] The law on all of this is basically non-existent. And there aren’t enough people who really understand the nuances who are also lawyers. It’s a whole mess which results in companies getting to decide for themselves. Not good.

[Critical Oil Theory Salesman] Hard agree. I’d imagine that we would see a completely different set of legal interpretations if the open source community trained a GPT3 model on Microsoft’s publicly available code.

Perhaps—that would be an interesting experiment. Is Microsoft really ignoring licenses? If not, Twitter is disseminating incorrect information. If yes, then Microsoft has designs on open source information in a way that outfoxes Amazon-type of open source maneuvers. But Microsoft is busy securing its own code and may want to envelope GitHub is the same cyber goodness.

Cynthia Murrell, July 22, 2021

Comments

Got something to say?





  • Archives

  • Recent Posts

  • Meta