John August

John

AI and Copyright

News, WGA

On Wednesday, I participated in a [listening session](https://www.copyright.gov/ai/agenda/5-17-Final-AI-AV-Session-Agenda.pdf) on AI and copyright hosted by the U.S. Copyright Office.

I was representing the WGA West. Other speakers included reps from the DGA, Public Knowledge, the Motion Picture Association and the National Association of Broadcasters.

Video will be up from the event in a few weeks.

Update: [Video is now available](https://stream-media.loc.gov/copyright/Copyright_and_Artificial_Intelligence_Listening_Session_Audiovisual_Works_05172023.mp4)! You can also read the [transcript](https://www.copyright.gov/ai/transcripts/230517-Copyright-and-AI-Audiovisual-Works-Session.pdf), and check out [all the other sessions](https://www.copyright.gov/ai/listening-sessions.html).

In the meantime, I wanted to share my opening remarks, which were a group effort with the team at the Guild.

—

My name is John August. I am a screenwriter and member of the negotiating committee of the Writers Guild of America West, a union that represents thousands of writers who create the content audiences watch every day in theaters, on television and on streaming services.

This is a unique moment for me to be speaking to this issue, because the subject of AI and its role in our industry is a major point of contention in the Guild’s ongoing nationwide strike against the major motion picture and television studios.

While writers who work under the Guild’s collective bargaining agreement are not copyright owners—we create works for hire—the Guild has negotiated an assortment of contractual rights in the works we create, including the right to residual payments for the reuse of our work across media platforms.

In the current negotiations, the Guild has made a proposal to regulate AI for the first time in our contract. The broad purpose of the proposal is to prevent our employers from using AI to devalue the work that writers do—to lower our pay, to deprive us of credit or attribution rights, or, in the most extreme case, to eliminate the need to hire writers altogether.
The proposal would also prohibit the companies from using material written under the Guild’s agreement to train AI programs for the purpose of creating other derivative (and potentially infringing) works.

The companies’ response has been telling: not only did they reject our proposal, they refused to engage on the issue at all. The most they have said is that the technology is new and they are not inclined to limit their ability to use this new technology in the future. This is an ominous response in the eyes of our members and one of many reasons that 11,500 writers have been on strike since May 2.

We often speak of copyright as a means of protecting works of authorship. But copyright was created with the intention of protecting **authors** from appropriation and theft. As we discuss the impact of AI, we need to remember the human authors and not just the corporations who employ them.

—

Following our opening remarks, panelists had the opportunity to answer questions about what rules and policies the Copyright Office should consider.

Speaking on behalf of the Guild, and its 11,500 members currently on strike, I focused on the need to protect the human creators of work.

You can read the [opening statement](https://publicknowledge.org/policy/opening-remarks-at-the-copyright-and-artificial-intelligence-listening-session/) of John Bergmayer at Public Knowledge, which specifically mentions the WGA strike. I’ll post other statements as I find them.

Which of these answers was written by AI?

Tools, Words on the page, Writing Process

[GPTZero](http://gptzero.me/) is a new service created by Edward Tian designed to sniff out AI-generated text.

I’m dubious, in part because I’m not sure I could reliably spot the differences between sentences assembled by humans versus machines. But maybe that’s just my human bias.

As an experiment, I decided to run a quick, not-at-all-scientific test. I asked ChatGPT to “write three paragraphs explaining what a manager does for screenwriters.” I also looked up the answer for [“What does a manager do?”](https://screenwriting.io/what-does-a-manager-do/) on screenwriting.io.

Here are the two responses, one written by a human ((Stuart Friedel wrote most of the early screenwriting.io answers.)), the other by an AI. See if you can spot which is which.

### SAMPLE A
> In brief, a manager manages your career.

> This includes tangible acts, like setting up meetings, giving notes and feedback, and assisting with deal negotiation, as well as the more abstract, like helping steer your career’s direction.

> Managers, unlike agents, are not required to be licensed by the state. They usually have fewer clients than agents, and often take more personal and less business-focused roles in their clients’ careers. Managers, are allowed to take producer credits on their clients movies, which agents are not.

> Like agents, a manager typically receives 10% of a client’s earnings. A manager should never ask for an upfront or flat fee.

### SAMPLE B

> A manager for a screenwriter plays a vital role in helping the writer navigate the film and television industry. They can assist with finding and developing new ideas, as well as helping the writer shape their existing ideas into marketable concepts. This often involves providing feedback and suggestions to help improve the story and make it more appealing to potential buyers.

> Managers also help screenwriters with the business side of things, such as identifying potential buyers and pitching the writer’s work to them. They can also help negotiate deals and contracts, as well as manage any legal issues that may arise. In addition, they can help the writer navigate the many different players in the industry, such as agents, producers, and studio executives.

> Finally, a manager can also help screenwriters with their long-term career goals by helping them identify opportunities for growth and advancement. This might include working on larger projects, transitioning to other forms of media, or developing a personal brand to increase their visibility and credibility in the industry. Overall, a manager plays a critical role in helping a screenwriter achieve success in the competitive film and television industry.

For each answer, GPTZero measured the text’s Perplexity — the randomness of the text — across several ranges.

| | Total | Average | Highest |
|———-|——-|———|———|
| Sample A | 41 | 136 | 219 |
| Sample B | 11 | 36 | 49 |

That 219 is an example of “Burstiness,” which GPTZero defines as “spikes in perplexity.” For Sample A, the opening sentence has a perplexity of 219, whereas highest perplexity in Sample B is just 49.

That’s the tell, apparently.

> Some human written sentences can have low perplexities, but there is bound to be spikes in perplexity as the human continues writing. Contrastingly, perplexity is uniformly distributed and constantly low for machine generated texts.

After crunching the numbers, GPTZero renders its verdict.

SAMPLE A (score 66.44):
Your text is likely human generated!

SAMPLE B (score 9.93):
Your text is most likely to be AI generated!

On both counts, GPTZero got it correct.

But how hard would it be to fool the system? I spent two minutes editing the AI-generated version into the following:

### Sample C
> Managers can play a vital role for screenwriters. Basically, they help the writer navigate the film and television industry. They can assist with finding and developing new ideas, or shaping their existing ideas. They provide feedback and suggestions.

> Managers also help screenwriters with the business side of things: who the buyers are and how to pitch to them. They can help negotiate deals and contracts, as well as manage any legal issues that may arise. They help the writer navigate the many different players in the industry, including as agents, producers, and studio executives.

> Good managers think about screenwriters’ long-term career goals. This might include working on bigger (or smaller!) projects, transitioning to other forms of media, or developing a personal brand. Overall, a manager can play an important role in helping a screenwriter achieve success in the competitive film and television industry.

| | Total | Average | Highest |
|———-|——-|———|———|
| Sample A | 41 | 136 | 219 |
| Sample B | 11 | 36 | 49 |
| Sample C | 19 | 106 | 439 |

The verdict:

SAMPLE C (score 118.77):
Your text is likely human generated!

## Some takeaways

– The ChatGPT answer was well-written and potentially helpful for someone who didn’t know anything about managers.
– GPTZero is looking for patterns a human likely wouldn’t notice, which makes sense. But an AI model trained to provide responses with high perplexity and burstiness would likely evade detection. It’s interesting to see this arms race play out, because adversarial networks are already a fundamental tool in AI.
– It’s currently very easy to fool GPTZero with just minimal rewriting. If this tool meant to detect AI-generated term papers, students will quickly learn how to evade it.

How ChatGPT resets the levels

Follow Up, Geek Alert, Hive Mind, Meta

On a [recent Scriptnotes](https://johnaugust.com/2022/rian-johnson-returns), I showed off ChatGPT to Craig and Scriptnotes guest Rian Johnson. We asked it to write up a plot summary for a future *Knives Out* movie set in space.

We got an answer similar to this query I just ran: ((The prompt is “Give me a plot summary for a Knives Out set in space.”))

> In the far-flung future, wealthy space entrepreneur Harrison Thorne is found dead in his luxurious space station, the victim of an apparent suicide. However, Detective Benoit Blanc, a renowned investigator known for solving even the most complex cases, has a hunch that foul play is at work. As he begins to investigate, he finds that every member of Thorne’s eccentric and contentious family has a motive for murder.

> As the detective delves deeper into the mystery, he must navigate the tangled web of lies and deceit spun by the suspects, all while trying to stay one step ahead of the cunning killer. With the help of Thorne’s loyal assistant Marta, Blanc must unravel the truth before the killer strikes again.

> As the body count rises and the tension mounts, it becomes clear that no one is above suspicion in this high-stakes game of space-age crime and deception. Will Detective Blanc be able to solve the case before it’s too late, or will the killer get away with murder in the cold vacuum of space?

None of us were blown away by the plotting. It felt like it was copy-pasting ideas from the first movie, with a lot of generic pitch filler. Still, we all agreed that the sentence-by-sentence writing was surprisingly good. It didn’t feel like it was written by a bot.

When we asked it to write a scene, it was functional but weightless. Here’s one based on the summary above: ((Prompt: “Write me a scene from this movie where Blanc first meets Marta, in screenplay format.”))

INT. SPACE STATION – DAY

Detective BENOIT BLANC, a middle-aged man with a southern drawl, walks through the sterile corridors of the space station, his face etched with determination. He approaches a door labeled “Harrison Thorne – CEO.”

BLANC

(to himself)

Time to get to work.

He opens the door and enters the office, which is spacious and well-appointed. A young woman, MARTA, sits at a desk, typing away on a computer. She looks up as Blanc enters.

MARTA

(stands up)

Can I help you?

BLANC

(flashes badge)

Detective Benoit Blanc, ma’am. I’m here to investigate Mr. Thorne’s death.

MARTA

(nods)

I see. Well, Mr. Thorne was a good man. I can’t believe he’s gone.

BLANC

(nods sympathetically)

I understand. Is there anything you can tell me about his final days, or any unusual circumstances leading up to his death?

MARTA

(hesitates)

I’m not sure. I mean, he was under a lot of stress with the business, but he always seemed to handle it well.

BLANC

(nods)

I see. Well, thank you for your time. I may have more questions for you later.

MARTA

(nods)

Of course. I’ll do whatever I can to help.

BLANC

(smiles)

Thank you, ma’am. I’ll be in touch.

Blanc exits the office, deep in thought. The investigation has only just begun, and he has a feeling that it won’t be an easy one.

To be generous, it’s not nonsense. Unlike [earlier attempts at algorithmically-generated screenplays](https://arstechnica.com/gaming/2021/05/an-ai-wrote-this-movie-and-its-strangely-moving/), you can follow what happens, and characters do feel like they’re talking to each other. But there’s no nuance or character motivation despite the endless parentheticals.

It’s not a scene, but a simulacrum of one.

Listener JR, who makes his living using tools like these, wrote in with an interesting perspective:

> Language models are built on “training data,” which is the text you feed into a learning process to produce the output. For very sophisticated models, the training data is vast: for something like ChatGPT, it includes something like _all the text you can scrape off of the last twenty years of the Internet, or so_.

> But this means ChatGPT is about as smart as the average writer on the Internet has been over the past twenty years — and indeed, the models that comprise GPT **drag the results toward the average, not the extraordinary**, because the average has much nicer statistical properties than the extraordinary for companies that seek to produce a marketable, scalable product from their models, which requires the ability to tweak, diagnose, and defend what you’re selling.

ChatGPT isn’t optimized for breathtaking originality because that’s not its goal. It’s meant to be conversational and approachable, not wild and daring. It’s giving you hamburgers rather than Japanese wagyu sando with white truffle shavings because 90% of the time, that’s what you want.

JR continues:

> Ultimately what these models mean is that with the click of a button you can now be just as good as the average writer who posts content to the Internet, and so the old “average” is now the new “zero.” **If you wrote at the average level of the Internet in 2022 you now write at the zero level.**

That’s a big idea, so let’s turn it over in our hands a bit.

Let’s imagine we can pick something that represents the average quality of English writing on the internet. It might be an article about [holiday lobster prices](https://news.yahoo.com/plenty-fish-sea-eaten-holiday-201000334.html) or a piece of [Peloton marketing copy](https://blog.onepeloton.com/fitness-accountability-partner/).

from yahoo news: Dec. 21—Maria Minella shopped at Schiff's Marketplace on Tuesday night to stock up for an upcoming family tradition. The Madison Twp. resident purchased smelts, scallops and smoked whiting from the Scranton store on North Main Avenue. They were a bit pricier than in years past, but that didn't matter much to her. "We like to do seven fishes for Christmas Eve," she said. "It's special for the holidays, so you pay a little extra to keep up with the family tradition." While prices of some fish are rising, the cost of other traditional Christmas Eve and New Year's Eve staples is dropping, according to area grocers.

screenshot from Peleton blog post: We get it: Staying motivated to work out regularly can be hard. But that’s where an accountability partner comes in. An accountability partner is there to cheer you on and encourage you to keep going, even on the hard days. They’ll drag you out of bed in the morning and keep you focused on your goals. And you’ll return the favor for them, of course—because accountability partners support and lift each other up. Looking to find an accountability partner of your own? We asked some of our Members to share how they found accountability partners and how that support has helped them achieve their goals.

Could any English speaker write at that level? No.

You probably don’t have to think that hard to come up with several people in your life who can’t write up to the standards of “Yahoo News article” or “Peloton blog post.”

Many folks don’t write, won’t write, can’t write.

In fact, if we want to consider it “median” writing quality rather than “average,” then approximately half the English speakers on the internet couldn’t hit this level of sophistication.

Only now, with tools like ChatGPT, they can. It’s very good at taking awkward, ungrammatical writing and making something decent out of it.

I fed ChatGPT this text —

> if youre stove won’t light here’s what you do 1) check that that the gas is on turn the knob do you hear it 2) lisen if it’s clicking that’s the igniter. if it’s not clickign you could light it with a match 3) lift off the burner cover cuz sometimes that the issue 4) try another burner

— with the instruction to “Rewrite this as a good blog post.” Here’s the result:

> If Your Stove Won’t Light: A Step-by-Step Guide

> If you’re trying to cook a meal and your stove won’t light, it can be frustrating and time-consuming. Here are some steps you can try to troubleshoot the issue:

> 1. Check that the gas is on. Turn the knob and listen for the sound of gas flowing. If you don’t hear it, the gas may be turned off or there may be a problem with the gas line.

> 2. Listen for the igniter. If you hear a clicking sound, that means the igniter is trying to light the gas. If you don’t hear a clicking sound, you can try lighting the burner with a match.

> 3. Lift off the burner cover. Sometimes, food or debris can accumulate underneath the burner cover and prevent the gas from lighting. Removing the cover may resolve the issue.

> 4. Try another burner. If all else fails, it’s possible that there is a problem with a specific burner on your stove. Try using a different burner to see if it lights properly.

> By following these steps, you should be able to get your stove working again in no time. If the problem persists, it may be a good idea to call a professional for assistance.

Imagine you’re a technician asked to document how a system works. You could describe the steps and let ChatGPT do the writing. Would you still need to proofread it, to make sure it’s not hallucinating something wrong? Absolutely. But most reluctant writers are competent readers. They should be able to spot when it’s gone off track.

Where I think these tools have even more benefit is for people who don’t feel confident in English. Type in your native language and let the AI do its work, translating not just the words but the intent.

Which brings us back to JR’s thesis, that a tool like ChatGPT resets the levels. In a world in which everyone can write at the old “average” level, what happens?

I’m not sure. There’s not a great comparison in recent human history. Calculators enabled ordinary people to do a lot more number-crunching, but were they transformative? (For the human calculators of *Hidden Figures*, perhaps so.)

A world filled with AI-polished text might push us in unexpected directions. To err is human, and we might find language incorporating mistakes to porve a hooman wrote it. Or we might use these tools like Instagram filters to make us look like someone we aspire to be.

Personally, I don’t think screenwriters are particularly endangered, at least not in the short term. Future tools may be trained and tuned for scriptwriting, but they’ll be working off a corpus of all the screenplays that have come before. **And that’s how it’s always been.**

JR notes:

> All writers are competing against the cumulative past of our species and its semi-automatic weapons of creative destruction, and are competing to stand out from an average defined over the past couple of decades or so.[…]
> Writers should only worry about automated writers as much as they are already worried about the agglomerated capacity and output of our species up until the present moment, which, eh, we can do better.

Social media has always been dying

Meta, News

For this week’s [Inneresting](https://inneresting.substack.com) rebroadcast, editor Chris Csont chose [this 2011 blog post](https://johnaugust.com/2011/all-yourselves-belong-to-us) in which I wondered how much of ourselves we present online. Even then, I was looking backwards:

> In the seven years I’ve been running the blog, I’ve noticed the online version of myself drifting closer to the “actual” version. But there is still a difference, and that’s deliberate. Even though this site has my name on it, it’s still a fairly controlled product: a ton of useful information on screenwriting. You’re getting the screenwriter John August, not the Eagle Scout, the cook or the Real World/Road Rules Challenge completist.

For the newsletter, I wrote up a blurb to put the post in context which ended up nearly as long as the post itself.

Eleven years ago, I was talking about Facebook, a platform I’ve largely abandoned. Nearly all the same points apply to Twitter and Instagram today.

Twitter has been problematic for years, with its trolls and mobs and twitchforks. But I’m honestly grateful for it. It’s where I first met many colleagues and learned new ideas, mostly by listening in on conversations from communities I wasn’t a part of.

For all the talk about Twitter leveling the playing field, allowing direct access to powerful people, I found my blue-check status to be incredibly useful. The moment I followed someone, they would get a notification and often follow me back. That digital introduction made it easy to DM them a question like, “Hey, would you ever want to come on Scriptnotes?”

Scrolling through someone’s Twitter gave you a sense of who a person was, or at least how they wanted to be perceived.

I’m talking about Twitter in the past tense because I think it’s dead. It may still exist in six months or six years, but Elon Musk’s erratic decisions have broken its hold on the conversation. It will ever feel essential again. Maybe that’s good. Other weird things can flourish in its absence.

Meanwhile, I still use [Instagram](http://instagram.com/johnaugust), particularly Stories and DMs. If I meet someone at a party, Instagram is probably how we’ll keep in touch. Yet it’s an every way a more extreme version of what worried me back in 2011. It’s curated, performative and artificial. It feeds body dysmorphia and parasocial relationships. But it doesn’t feel like it’s contributing to imminent collapse of democracy, which in 2022 is the low bar we’ve set.

As I’m posting this, I haven’t officially left Twitter. I still skim and reply. But I’m not going to be engaging in any discussion or debate. I’ll certainly try out Hive and [Post](https://post.news/johnaugust) and [Mastodon](https://mastodon.art/@johnaugust), just like I used to have Tumblr and others. But I don’t expect any of them to replace Twitter for me.

Just as it was 12 years ago and seven years before that, I suspect this blog will remain the official version of me, so follow me here. [RSS forever](http://johnaugust.com/feed).