The First Anti-AI Strike: The Writers Guild of America Protests AI Writing

0:00This episode of the AI Breakdown originally appeared on YouTube on May 3rd. In it, we cover the Writers Guild of America strike and what it says about how industry might fight back against AI. Before that, however, we cover the day's top headlines, including Samsung banning generative AI, TikTok adding an AI label, and a new ChatGPT competitor. Welcome to the AI Breakdown Brief, the AI headlines you need to know today in five minutes or less. Step back, ChatGPT, there is a new contender in town. Well, sort of. Inflection AI has just launched its new personal AI called Pi. Now, this is a company that's raised something like$285 million.

0:43It shares a co-founder with LinkedIn and Reid Hoffman, as well as another co-founder of Google DeepMind. So it's got a lot of clout behind it. The idea of Pi is that it's supposed to be more personal. The CEO, Suleiman, says it's very balanced and even handed on political issues or sensitive topics, but it can also sometimes be funny and creative. This is what it looks like. And part of what makes it different or is trying to be different is that it remembers context from your previous conversations. And in some ways, it feels much less like a competitor to a chat GPT, which is helping people code or write copy, and instead something akin to the call Annie app that we demoed on this channel a few days ago.

1:23Now, whether people actually want that out of their AI remains to be seen. Next up, TikTok is considering a new content label for AI generated content. This is becoming more and more important as AI content is on the rise. Now, part of this is driven by things like the absolutely viral Drake tracks that have been coming out and flooding all over TikTok. But part of it's in a response to fake photos like this one that went viral from Selena Gomez at the Met Gala when Selena Gomez didn't go to the Met Gala. I think you're going to see a lot more companies take steps like this AI consideration in the months to come.

1:58Now, bringing it back to ChatGPT, there is a new plugin called Code Interpreter that a lot of people are getting really excited about. The neuron here says that people are saying that Code Interpreter rivals junior level data analysts. Basically, this allows you to upload a CSV file and the interpreter then, one, provides suggestions for visualizations and analysis, but then actually does them. So an example that they give is someone uploaded a CSV of San Francisco crime data and got bar charts over time, heat maps of distribution, scatter plots of incidents. This is a pretty amazing tool if you think about the way that data can be visualized for more high impact sharing.

2:38And while this might be worrying for data analysts who are seeing their job prospects change before their eyes, it could be a democratizing force for how data can be useful in society, which in and of itself demonstrates the good and the bad, the challenge and the opportunity of AI in one small little example. Speaking of the challenge of AI, Samsung has banned staff use of generative AI after a data leak into chat GPT. So this data leak happened last month and Samsung got very concerned about it. And this is part of a larger trend of companies being concerned with their proprietary data being fed into chat GPT.

3:15Now, of course, OpenAI has announced that they have a forthcoming business version of ChatGPT that tries to address this problem by having it be private by default. They also just announced a private mode for everyone else as well. So OpenAI is trying to solve this type of issue. But for the meantime, Samsung has banned generative AI in the workplace on workplace devices. Finally, yesterday we discussed how the White House is taking a bigger and bigger interest in AI. and that seems to be validated in the news that Vice President Kamala Harris is set to discuss AI in a meeting with the CEOs of companies including Google, Microsoft, OpenAI, and Anthropic on Thursday.

3:53Now interestingly this doesn't appear to be a press event type of meeting because it wasn't announced by the White House. It was something that CNBC got their hands on the invitation. The CEOs who have been invited aren't commenting on it so hopefully they actually have the chance to have a productive discussion. That's it for today's breakdown brief. I'll see you back here for the full breakdown a little later today. Today on the AI breakdown, we're talking about the Writers Guild of America strike, which has AI and AI replacement and automation right at the heart of it and asking what it suggests about the state of the conversation around AI disruption in industries across America and the world.

4:32Yesterday, the Writers Guild of America strike began. And there are many reasons behind the strike. There are questions of wages and lowered prices and use of day writers and all these sort of things. But right at the heart of it is fears of AI and a concern about what Hollywood executives and studios might use AI for in the future and how it might impact the people who actually write the shows and TV and movies that we watch. Now, specifically, the Writers Guild of America is in contract discussions with the Alliance of Motion Picture and Television Producers, and they're focused on what they call their minimum basic agreement, or MBA.

5:13In those discussions, here's how they framed artificial intelligence. What the Writers Guild asked for was to regulate use of artificial intelligence on MBA-covered projects. AI can't write or rewrite literary material, can't be used as source material, and MBA-covered material can't be used to train AI. That was rejected by Hollywood executives who countered by offering annual meetings to discuss advancements in technology. The fear is, of course, that basically executives will create a new process for writing TV and movies by which AI might be used to come in and do the basics of screenplays.

5:49And then Hollywood would hire writers at day rates to punch up or make those scripts a little bit better. Now, is this actually what's going to happen in Hollywood? There are kind of a variety of takes. On the one hand, you have the comical takes like Mike here who wrote a fake episode of Succession using fake AI that's just terrible, obviously making the point that AI can't do this. And of course, a lot of Hollywood writers feel very similarly. At the same time, Sarah Myers West, who's the managing director of the AI Now Institute, which is a research nonprofit that studies social consequences of AI, says that, quote, writers in particular are among the most likely to be affected.

6:55West continues, feeling. She writes,

7:07Now the question, of course, I had is holding aside the strike's justification. From an efficacy standpoint, doesn't it create the perfect incentive and opportunity for studios to actually go try the exact thing that writers fear. General Catalyst investor Zach Kukoff says something similar, writing, Strike breaking becomes much easier when AI can be a 24-7 scab working for pennies on the dollar. Tough time to be a creative worker like the Writers Guild going into a strike. But WGA is smart to negotiate this now. Their position only gets worse as AI matures. And indeed, that's also a lot of the response that I got to my Twitter question as well, that sure, maybe Hollywood might want to do that now, but the technology isn't really at a point yet where it's viable.

7:53And so this is exactly the right time to be negotiating exactly this. Adam Conover, who created Adam Ruins Everything, said, In terms of companies using AI in order to break the strike, I'd like to see them try. It's not going to work. It's not easier to replace us with AI than it is to find someone to write the scripts. And that's not possible for them to do because it's an extremely skilled profession. Now, of course, the problem with that thesis is that it doesn't necessarily matter if Adam is right, if executives are willing to just churn out crap. But this, of course, isn't just about the Writers Guild.

8:26Justine Bateman here tweets,

8:39And this, of course, broadens out the conversation to more generally what AI is going to do to industries. I saw a very interesting and kind of glib, cocky even tweet from Jan LeCun, who's the chief AI scientist at Meta yesterday. He retweeted something from Benedict Evans that suggested that the AI discourse about jobs destruction is way out ahead of itself. And Jan added, Every economist I know says that it takes 15 to 20 years before a new general purpose technology has a measurable effect on productivity. The delay is determined by how fast people learn to use it. So no, AI is not going to cause instant mass employment.

9:16It's going to displace jobs over time and make people more productive, just like every other technological revolution before that. Now, there are a lot of challenges with this. One, I will say, is just the confidence and lack of humility inherent in the statement, assuming that you can apply the model of previous technology changes to AI. But the second is, it's just not exactly being borne out in reality. Take, for example, news from earlier this week that IBM is replacing 7 ,800 jobs that they would have hired for in areas like human resources and back office support with AI and automation. Dropbox announced that they're cutting 500 jobs in their pivot to AI, not because they can automate those jobs away, but because the skills aren't relevant for what they need to build for the future.

10:02So between IBM, Dropbox, and now the Writers Guild of America, I don't think that dismissing the concern around AI's disruption to industries and jobs is quite so easily done. Now, where Jan is right is that one of the impacts of AI will be increased productivity for human workers as well. Planet Money just published a long article about a team of researchers who have been using an earlier version of a chat GPT-like AI to help customer service workers, and they found that the hybrid AI human workers were much more effective, much more productive. The study found that they were able to resolve issues faster, to resolve more complex issues than were previously possible.

10:41That customers who had interacted with the hybrid human and AI customer service reps were more satisfied. So yes, there will of course be those productivity gains as well. And when it comes to the net impact, I think it is still really unclear. On the one hand, people are concerned about there being technological haves and have-nots based on these new AI platforms. But on the other hand, people are talking about how entire categories of jobs that currently enjoy a wage premium might no longer enjoy that wage premium because of AI. That would have the net impact of decreasing inequality, but not by raising people up, but by bringing an entire category of people down, which creates its own set of problems.

11:18The point of this is that this conversation is now here in a big way. The Writers Guild of America is probably just the leading edge of something we're going to see more and more. Technologists have long assumed that the forces of technological progress march on unimpeded no matter what. But I think that when it comes to AI, for the first time, society is going to ask some real serious questions about that. Engaging in good faith with that conversation, I believe, means probably not bringing all of our priors to it. I certainly don't know what the right answer is and where society nets out as better than it is now.

11:50I know that these tools are powerful and that I'm excited for them. I also know that the disruption could be hugely problematic. But you better believe that this is the beginning, not the end of the conversation. And more and more people are going to get involved, probably in a way that asks, is this really what we want? Anyways, guys, lots of food for thought. If you were enjoying the AI breakdown, I would so appreciate it if you would take the time to subscribe to it or go subscribe to the podcast version of it. Until next time, guys. Peace.

12:30Thank you.