Business Musings: WGA Agreement Regarding AI

Business Musings Current News free nonfiction On Writing

To get a free audio version of this post, click here.

Well, I’ve been in the weeds tonight. Thank heavens I’m heading to Las Vegas’s Pride Parade in an hour or so. Because my brain is quite mushy.

I just have read the entire Memorandum of Agreement for the 2023 WGA Theatrical and Television Basic Agreement. These are the terms that the Writers Guild of America’s membership finalized on Monday.

It’s really not fair to say I’ve read the entire agreement, because much of it is based on previous agreements that I don’t have easy access to. Besides, this memorandum is 98 pages long, and I really don’t want to (re)read all of the preceding agreements that this one modifies.

Still, I read it so you don’t have to.

There’s a lot of interesting material in here, things that I need to read carefully since I have options and deals that were more or less on the table when the strikes started. I do need to understand all of the terms of the MBA, and I’m simply not going to go that deep into the agreement on a Friday night when there are rainbow flags awaiting.

So what I mostly looked at was the brand-new section on generative AI.  That’s what I’ve been waiting for. When I initially read the reports on the deal, which did not include the details of the deal because it hadn’t been released yet, I felt that the WGA did the minimum on generative AI or what, in this agreement, is listed as GAI. (It sounds like something a cat would hork up.)

The early reports made it sound like the WGA got some of GAI right, and whiffed on the whole training the chatbot thing. But it’s more complicated than that.

First, let’s talk a little bit about the agreement itself. It seems that by all measures, the WGA was right to call their deal “exceptional” in their initial letter to the members.

This strike was vicious—to the writers, to the other guilds, to the support communities, to the studios, and to the City of Los Angeles itself. According to the Bureau of Labor Statistics, the dual strikes (writers and actors) have cost the entertainment industry 45,000 jobs since May…and counting, because SAG-AFTRA is still on strike at the time of this writing. (The actors union, for those of you who live under an acronym-free rock.)

Right now, the estimated cost to the U.S. economy is $5 billion dollars, but that will probably go up. (And it doesn’t count the impact on other countries, like Canada, where a lot of U.S. productions are shot.)

Because of all this pain, the deal had to be exceptional. People lost their homes. Many couldn’t feed their families. As I told a friend who had a “we’ll see” approach to the word “exceptional,” if the deal hadn’t been amazing, then no one would have trusted the WGA negotiators again.

The WGA got almost everything it wanted, maybe because the studio heads—who finally showed up at the bargaining table with a serious attitude about negotiating—realized that the writers were resolute. Initially, the studios thought the writers would give up when they started feeling the financial pain. Instead, the writers became even stronger.

Some of the reason for that was the fact that most of these writers had already experienced a major drop in income over the past five years, because the studios and streaming services used every device they could to stop paying writers what they were worth.

This strike, for the writers—and, really, for the actors—is existential. If they didn’t get what they needed, then they wouldn’t have careers. And no one else would be able to have a career either. It took the studios a long time to catch a clue—and they had to essentially override their negotiator—and by then, the damage to the industry was profound.

If you want to understand how negotiation is done and how to succeed at it, then study the narratives from the Writers Guild that will come out of this. The Hollywood Reporter has a good initial discussion with the WGA negotiating team. I’m sure there will be more articles like that one once the SAG strike ends as well. And, I’m convinced, the history of this agreement will be taught in college classes on bargaining, negotiation, and labor years down the road.

Most of this, though, only has a peripheral impact on you prose writers reading this. Unless you get some kind of movie or TV nibble, not much applies to you. (And if you get one, you better understand what the WGA does and what you should be offered. Most writers don’t. I do have a book that can help you, as a start anyway.)

There’s a lot of important things in the agreement on GAI. You’ll find those things in the Article 72 section on page 68. First, the WGA and studios agreed (and this is big) that GAI cannot be used indiscriminately to “write” screenplays or even stories that screenplays are based on. Why? Because the agreement follows the Copyright Office’s ruling that GAI has no human component, and therefore can’t be copyrighted.

There are strict rules now as to how GAI can be used in the writing of or development of screenplays. Those rules protect the writers quite a bit. The rules also protect the studios, because everyone wants to hold a copyright on a screenplay rather than lose it. The agreement outlines some sample uses, but I’m sure there will be others and some might have to go to arbitration.

Article 72.D makes clear what’s at stake:

A writer will be required to adhere to the Company’s policies regarding the use of GAI (e.g., policies related to ethics, privacy, security, copyrightability or other protection of intellectual property rights). Any purchase of literary material from a professional writer is also subject to such policies. A writer must obtain the Company’s consent before using GAI. The Company retains the right to reject the use of GAI, including the right to reject a use of GAI that could adversely affect the copyrightability or exploitation of the work.

The emphasis here is mine. “Adversely affect the copyrightability or exploitation of the work” is no small stuff.

But here’s the part that I initially thought the WGA whiffed. It was on the future of GAI. It’s Article 72.F and G:

The parties acknowledge that the legal landscape around the use of GAI is uncertain and rapidly developing and each party is reserving all rights relating thereto unless otherwise expressly addressed in this Article 72.…

[G] Each Company agrees to meet with the Guild during the term of this Agreement at least semi-annually at the request of the Guild and subject to appropriate confidentiality agreements to discuss and review information related to the Company’s use and intended use of GAI in motion picture development and production….

This is what I’ve been saying all along. The legal landscape around AI is constantly shifting. The contracts for the WGA (and SAG-AFTRA) had to be renegotiated in this changing landscape because the previous contract was ending, so they had no choice but to put in something. They negotiated what they knew—preventing the threat to writers’ jobs and the copyright as it exists right now—and saved the rest for later.

Notice that the companies will be meeting with the WGA on a regular basis to handle the shifting copyright issue during the term of this agreement (which is three years). A lot has changed since January. I can’t imagine what it will all look like in 2026.

There’s quite a bit in The Hollywood Reporter article on this part of the negotiation. I suggest you read it. I’m going to excerpt this part here, though. Union negotiating committee co-chair Chris Keyser mentioned they got two types of agreements on AI. One is particular to the industry as it stands right now. But on the tech going forward, he says:

instead of trying to negotiate beforehand a world we don’t understand, we retained every single right we have to negotiate for writers, and writers retain every single right they have both under the law and the MBA to protect themselves in circumstances of the companies using our material to train.

The critical phrase here is “instead of trying to negotiate beforehand a world we don’t understand…” That’s key. That’s what I’m trying to get you writers to do.

The negotiators here took the changing situation seriously, and understand how delicate the future might be. So many novelists are not doing either of those things. I recently saw a project from a writer I respect who is going blithely forward into seriously uncertain territory on a book project. I’m worried that this writer’s copyrights might be affected. But, as I tell my writer-students, you are responsible for your own career…

In other words, it’s not my job to police all of you. I do occasionally warn you in this blog just like I did last week, but I’m not in charge of what you do. (Nor do I want to be.)

So, that’s the WGA agreement. I’m curious to see what SAG-AFTRA negotiates, because they have even more issues than the WGA does. And, of course, there are all of those court cases I mentioned last week, along with several others going on. The September 27 Hot Sheet by Jane Friedman (requires subscription) has a good list of the class action suits going on right now, and some good analysis as well.

Okay, that’s it for me at the moment. I have a parade to attend. I’m going to move off AI for a while after this, unless there’s some incredibly huge development that I can’t ignore.

Do keep following all of the developments yourself. It’s important and will have an impact on you. What you don’t want is to be caught flat-footed, such as having your books removed from a platform because you got caught up in the AI hype and didn’t investigate changes to some Terms of Service.

It’s a risk. But then, if you’re conservative in your use of the new tech, you will minimize risk. It’s not always necessary to jump on a bandwagon, no matter how cool that wagon is.


The observant among you have noticed that my website is slowly being updated. Whew! That’s so necessary since I was using an eight-year-old theme. I also am going to make some other changes slowly across the site as the months go by, trying to bring it into the 2020s before we hit mid-decade. So thank you for your patience on this. I hope the changes won’t have too much of an impact on readability as we’re in the middle of all this.

And, back to our usual announcements.

I mentioned above a book that I wrote that will help you with any movie or television contact. The book is called Tips About The Film/TV Industry For Novelists. You can order it here.

The Bite-Size Copyright class that we offer will take some of the mystique off the way that copyright works. If you’re having trouble understanding copyright or if you just want a refresher, I recommend it.

And the weekly reminder: This blog is reader supported.

If you feel like supporting the blog on an on-going basis, then please head to my Patreon page.

If you liked this post, and want to show your one-time appreciation, the place to do that is PayPal. If you go that route, please include your email address in the notes section, so I can say thank you.

Which I am going to say right now. Thank you!

Click to go to PayPal.

“Business Musings: The WGA Agreement Regarding AI,” copyright © 2023 by Kristine Kathryn Rusch. Image at the top of the blog copyright © Andrew Genn.

9 thoughts on “Business Musings: WGA Agreement Regarding AI

  1. Thank you for providing a detailed and insightful overview of the 2023 WGA Theatrical and Television Basic Agreement. Your commitment to reading through the 98-page memorandum and distilling its key points for the benefit of others is commendable.

  2. One thing that most creators are not (understandably) tracking about LLM/GAI/etc* is the grift.

    Just like NFTs and crypto before them, large language models are only the latest tech grift/pyramid scheme and, like NFTs, creators are simply collateral damage.

    The good news: A.I. is costing more to produce and is proving harder to monetize in a way that will feed a pyramid, so it may crash and burn faster than previous tech bro grifts

    As to the concern, Kevin McLaughlin, that “It is not possible to detect it in a provable way. There is no way to prove a written work did or did not use AI, especially after human editing”, that’s actually going to cause creators MORE problems than many yet realize.

    Because programs to detect the use of large language models in writing are available. And, they’re being used to accuse people of (and penalize them for) using large language models, even when that’s false (see students “caught” using it and being accused of plagiarism — Jonathan Bailey of Plagiarism Today has been following this particular aspect at

    *It should be repeated, endlessly, that nothing about what people are calling “A.I.” is artificial- (since it relies on humans forced to work for slave wages) nor “intelligent” (since an algorithm cannot by definition be self-aware and thus “intelligent”).

    1. Nicely said. Agree on the monetization bit. Right now, the open source LLMs are expected to surpass GPT4 my sometime early next year. At that point, monetization of GPT4 becomes next to impossible, since anyone with an M2 MacBook can install their own open source LLM. Of course, by that time GPT5 will be out, but for many purposes GPT4-equivalent quality is sufficient. The speed with which open source AIs are catching up to (or in images, leading the way, even!) commercial ones is impressive.

      And you’re right about the so-called AI detection apps already causing problems. Colleges across the US were all busily warning their professors last August that they CANNOT use those apps to detect AI work, because if they try to fail a student on that basis, the school will lose the lawsuit which follows, and lose BADLY.

      Unfortunately, high school teachers haven’t as reliably gotten the memo yet, so I expect a few school districts will get sued for ‘detecting AI’ on a student’s paper and not being able to prove it.

      The good news is, once we get even a couple of cautionary tales, it’ll probably stop happening.

      And definitely NOT intelligent. 🙂 We’re nowhere close to general AI. Thank goodness, TBH!

  3. The WGA did an amazing job. 🙂

    From my POV, the real glory of their negotiation stemmed from the fact the WGA actually understood a CRUCIAL component of AI generated text:

    It is not possible to detect it in a provable way. There is no way to prove a written work did or did not use AI, especially after human editing.

    Because of that, they did t try to stop AI use.

    The WGA just said “AI or not, you still have to hire the same number of writers; those writers still get precisely the same credit with or without AI; and those writers still get the same pay, with or without AI.”

    This wipes out the WGA’s main concern of AI use: lost jobs.

    Super smart move, and it demonstrates a level of AI understanding I find rare among creators.

    The need to reassess as things develop is key as well. Like, we all know AI work DOES already have full copyright protection in the UK and Japan. The EU is considering whether to follow suit, and the USCO will almost certainly reverse course within two years as well. (Because the U.S. Copyright Office is already issuing copyrights to any AI work sent in, because they can’t tell what is AI and what isn’t – no one can.)

    So the landscape WILL change. That’s about the only sure thing. Setting up to change with it is wise. 🙂

  4. Is there any way you can set the emails to come from, say, Kris Writes (or the previous Kristine Kathryn Rusch), rather than There are zillions of WordPress sites.…

Leave a Reply

Your email address will not be published. Required fields are marked *