AMT Lab @ CMU

View Original

Let’s Talk: AI Policy and Artist Advocacy

What happened to the AI bill SB-1047? And how can artists advocate for themselves amid technological changes? In this Let’s Talk episode of the Tech in the Arts Podcast, AMT Lab’s Chief Editor of Research, Hannah Brainard, and Lead Researcher, Ian Hawthorne, dive into the recent developments in AI policy and its impact on artists.

Show Notes

See this content in the original post

Transcript

Hannah Brainard

Welcome to this episode of Tech in the Arts, the podcast series of the Arts Management and Technology Lab at Carnegie Mellon University. Our goal is to connect you with innovation happening in the field. And in this episode, we have a lot of exciting things to talk about. 

My name is Hannah Brainard. I'm the chief editor of research here at AMT Lab, and I'm joined by Ian Hawthorne, our lead researcher. Today we were planning to talk a little bit about AI, policy, the implications for artists, artists role in advocating in AI and the role more broadly in democracy. We have an election coming up, I guess, huh? So yeah, let's dive into it. Last episode, and I don't know if you listened to it with Dr. Crawford.

Ian Hawthorne

I did. It was good. Yes.

Hannah Brainard

We talked a lot about the California State Bill 1047, which was a very landmark bill, high in the headlines over the last month. And there was a lot of discussion. So at that point, it had been approved by the state Congress and was on the desk of Governor Gavin Newsom to either sign or retold the bill. And that's where we left it. But there's an update. He did not sign the bill. He vetoed the bill.

Ian Hawthorne

R. I. P.

Hannah Brainard

So, yeah. That's kind of an interesting position that we're in. It passed in the state Congress. It was vetoed by Gavin Newsom, but he released a statement with some of the reasons as to why it was vetoed. And I was a little bit surprised by some of those reasons. I mean, I don't know sort of what you thought. 

Ian Hawthorne

Yea, I was definitely. I would put it this way. I wasn't surprised but I was disappointed. And that reveals a little bit of my own personal stake in how I thought this should have gone. So, he basically rejected it and said it's too nonspecific. It doesn't account for AI programs used in high-risk cases with super sensitive data. And one of the biggest things that struck me was that he basically said it only targets large models. As a refresher of the policy stated that large models, if they were big enough, needed to have regulations in place, they needed to have a safety audit, and they also needed to have a kill switch in case they ended up leading to, you know, mass death or bio warfare, which I think is reasonable.

But one of his biggest retorts was basically, you know, well, it's only targeting the big guys. What if there's a super dangerous small guy, which is a fair response, but I, it does beg the question of, well, if that's the case is nothing better than something. So that was definitely one flag that was raised for me.

The other thing that it would have done was establish a state body to regulate these going forward, which I think for some of the big regulation supporters is probably the bigger disappointment there. Just the fact that this really would have been the first major policy to pass anywhere in the US that does anything to regulate AI and I think using it as a.. 

Hannah Brainard 

You know, using it as like a starting point,

Ian Hawthorne: 

Yeah. A model. 

Hannah Brainard

California in particular, exactly. Yeah. And California is home to, it's had 32 out of 50 of the top AI companies in the world. So, this is not unprecedented that California would create some kind of legislation like this that would become a model, you know, within the United States and internationally, even this is a landmark sort of bill.

Ian Hawthorne 

It's fitting, and I wonder, I would wonder if part of his worry too, beyond the influence of larger AI companies in politics, in elections, and potentially in donating to him, and also supporting parts of the California economy, is I would worry about sort of the Tesla effect, where if you start to regulate us, we can simply move to Texas.

So, I would wonder how easy that is to do with servers and data farms and, all of that infrastructure for AI companies, but I'm sure that would be a possibility that was probably threatened behind closed doors at some point.

Hannah Brainard 

And, you know, his statement said that this isn't a closed book. They are planning to continue to investigate what the best model would look like to regulate AI. I have to pull a quote here from his statement, "Ultimately any framework for effectively regulating AI needs to keep pace with the technology itself." So, you know, I guess continue research and see what happens next.

Ian Hawthorne 

Yeah. And I, I can't disagree with that either because obviously it's better to have more nuanced policy than blunt force policy, but there is a question of if we're going to keep reiterating our regulations, when are we actually going to pull the trigger to, to do something? So, it's back up in the air, the, the policy is not going to happen. I took my moment of silence for it earlier, but we're back to square one for a lot of people.

Hannah Brainard 

And, you know, one of the biggest groups that, AI policy is sort of impacting. So, there were a number of other bills in front of Gavin Newsom, specifically AI related. This was probably the biggest, but lots of implications and specifically implications for artists in terms of how this regulation goes.

Even just this month, there've been some really interesting stories about how AI is being used in different creative work, and what that could mean for creators. One that you were talking a little bit about before, Squarespace, could you talk about that. What are they doing?

Ian Hawthorne 

Yeah, absolutely. So one of the big announcements that Squarespace made just this past month was the introduction of AI into their website design interface. And it's basically allowing you to use text generation to better curate how you end up designing a website. So instead of picking from premade templates, they have sort of a curation model built into the technology that you can interact with like a LLM type of engine.

And they took a little longer to get here than other places did. They actually made a comment that other website builders rushed into the AI space, and they really wanted to wait a little longer because they're a very big player, and they wanted to integrate it well. And it seems like from a lot of the examples they've put, it is a little more thoughtful than other web builders. 

And, and they're really, it's really, it's funny because they're really arguing it on quality. Squarespace is basically saying we have better quality AI. We can make prettier websites with it because they have a ton of digital assets too. And it, even though it is trained using AI, it feels bespoke. And that's sort of the offering that they're giving to people is that it allows you to really build your own bespoke website, even though you're using a model, for instance.

Hannah Brainard 

Well, that's a really interesting aspect of it. You know working with so many different smaller nonprofits in particular, the ability to make some of those edits to your own website independently would be a huge benefit.

Ian Hawthorne

Definitely. Yeah. And it's, it's always a give and take, right? Because on the one hand, it's like, this is great for nonprofits, small budgets. You can finally take a lot more control over your website. On the other hand, you don't end up contracting with the web builder. So, I do think that this is probably a net positive in the end because it just increases access to that knowledge and that ability for a lot of people. But that was, it was definitely just striking to see it. To see AI encroach in that realm, among all the other realms that it's going into, it really is starting to be everywhere at this point. 

Hannah Brainard 

That's true. And another one that I'm almost hesitant to say it here on our podcast. Google Notebook LM went viral this month. So it's originally created for the purposes of research and summarizing large sets of data. So you can have up to 50 different sources if it's documents or websites or videos or audio clips, slideshows even. All sorts of different types of files that you can input and then get an overview of the material, and an audio overview. So, one went viral this month. It sounded, you know, fairly realistic and like summarized information pretty quickly, which is interesting.

 Ian Hawthorne 

Yeah, definitely. I think the spoken element of AI is becoming really fascinating very quickly. You're looking at these types of programs like notebook LM and people starting to make podcasts with it, where you can basically feed it entire text strings and it will create very realistic sounding, male/female vocals to narrate it. Which to all the viewers out there should cast some doubt on whether or not we are AI right now. Disclosure. We're not, but... you may not know.

Hannah Brainard 

I was reading a Wired article about it, and it was talking about how if you are a podcaster, you're probably not in the place to start worrying just yet, because I think there's something to this, like, human element. For example, saying the relationship that you form with the people that you're listening to, whether you know them or not, there's like something to that, that people aren't ready to give up yet. The technology is still not quite there, that it's replaceable. It also doesn't have the ability to use your own voice to create audio yet, so, who knows, in the creation process, what will happen next, though?

Ian Hawthorne 

We are irreplaceable and that makes me happy.

Hannah Brainard 

Agree, I agree. 

Ian Hawthorne 

We are perfectly imperfect.

Hannah Brainard 

So true. Well, another example, music teachers teaching piano. Talk about that. 

Ian Hawthorne 

Yeah. So this is kind of another inspiring one where it's a company ROLI Airwave is now able to use AI, paired with hand tracking to teach piano to people, which is a very sort of simple innovation. It's almost one that I read, and I was like, I'm surprised this hasn't happened yet before. But it just adds in AI to all the other sort of tech interventions in arts education.

And again, what seems fairly simple and straightforward does to some degree eliminate the role of a piano teacher. And again, it always brings it back to this question with technology of access over employment, right? Where all of a sudden more people could be using these programs to learn piano. And all of a sudden we have, you know, 10 times the amount of wonderful concert pianists that we can employ and that can [perform] their music. But at the same time, you know, it just makes you wonder,

Hannah Brainard 

It's so true. Like, having this access to more people to be able to learn music is so powerful. But the people who earn their livelihoods in teaching piano and that human experience of like, sitting down on the bench with a piano teacher or those little, you know, numbers and colored stickers as I remember way back. Yeah, so there's something to that balance of, you know.

And I think that kind of leads to another discussion of the artists that are impacted. How they can really advocate for their position and how they're affected by AI in these different places. How do you see arts involved in advocacy?

Ian Hawthorne

So, I see arts involved in advocacy very deeply. I think that, and this is my opinion, I know other people might differ on this, but I do think that the arts are sort of the fundamental mouthpiece of a culture or a society. You know, this is where messages can be made that are uninhibited. And I think that a lot of artists naturally tend towards the political.

I think that's been the case for quite a long time now. But you're seeing with technology over the past few election cycles, a lot more artists and artist collectives starting to create things like “get out the vote” campaigns. And hosting, sort of, charity auctions for works that may or may not have a political or an Americana theme to them in order to fundraise for 501C3s that are aimed at getting out the vote, aimed at voter participation, advocacy for certain issues.

Artists and artist collectives are freer to be a little more political than an arts institution might be. Anybody who is aware of 501C3 law knows that there is a limit to the amount of political lobbying that they can do. That is to say, “get out the vote” and voter registration events, and even organizations acting as polling places, which I write about in one of my more recent articles, none of those are political activities when you think about it. They are pro-democracy activities that support government functions, but they do not make a statement on who you should vote for, which is why I think it's really interesting to see this growing trend of, it's usually museums acting as polling places in various cities across the country from New York to California, but also in between. We see it in Michigan. We see it in a lot of other states, too. And it's small but growing. I've noticed from year to year, you know, those who volunteered in 2020 are still acting as polling places in 2024. The Brooklyn Museum is a really great example of this. And they're becoming these sort of community spaces that are serving a purpose beyond just being a museum.

It's basically saying to your community, we want to help support the government that gives us a right to be here for you, right? That gives us the status as a 501C3 and that's sort of an important sub-mission of a lot of museums. A lot of museums, nowadays, there's this question of neutrality on a lot of issues. 

But when you really think about it, a museum may have the mission to spread a certain art form and increase access to an art form to a community. But I think a lot of them have undertaken this extra mission to some degree of supporting democracy of supporting community of really being a space of access and equality.

And I think that acting as a polling place or participating in these voter registration events, these “get out the vote” campaigns, that would be great partnerships between institutions and artists who are running them. All show the community, basically, what you stand for. 

Hannah Brainard 

Yeah, and not even like on a political level, like who you stand for what issues you necessarily stand for, but just that you stand for your community represent your community. It's a place for civic engagement. Yeah, that's really interesting. And not only does that like benefit the community, it also gets people in the doors of your institution which is ultimately beneficial for all.

Ian Hawthorne 

Exactly. It's a win-win. And I think, too, it is very important because we've been talking a lot about AI. We've been talking a lot about artists, art teachers, the arts being impacted by AI. They were some of the original sort of stakeholders in these arguments when you look back at the emergence of Dall-E 2. When everybody was making those artworks, those NFTs, issues of IP copyright, and even writers with, you know, way back when just chat GPT was the only AI people were talking about, the arts really have a central position to play in this discussion. And I think it's really important for artists to be aware of that, but also to know how to advocate. And so that's where I think participating in these kinds of “get out the vote” campaigns. Really trying to reinforce democracy in order to create a voting environment and a participatory environment where people are able to express concerns over the rise of AI, because it's not just affecting artists.

Now it's affecting website designers. It's affecting podcasters. It's everywhere. And it's, it's really just important that artists stay in that conversation and continue to advocate for these types of policies, like SB1047, that we were just morning at the beginning of this podcast, because there will be more going forward, you know, on state levels, hopefully on federal levels, and it's going to be really exciting to see where that regulation goes.

Hannah Brainard 

Yeah, absolutely. We're very close to a significant election, and I think, when we think about it, we think of the federal election, but in this case we're talking about at the beginning is a great example that it happens at all levels. This legislation affects artists, affects communities, arts institutions, so the role that they play in creating this democracy in the United States is just so pivotal.

So I highly encourage everybody to go check out that article on AMT-Lab.org that Ian wrote. Really insightful about this relationship in the community. 

Ian Hawthorne

Absolutely. Thank you.

Hannah Brainard 

Yeah, well, this has been a really fun discussion for me. I really appreciate it talking about some of these important headlines.

Ian Hawthorne

Likewise, I hope everybody who's a U. S. citizen and hears this, votes over the next two weeks. This might get dated very quickly because of that, but it's worth it.

Hannah Brainard 

Absolutely.

Ian Hawthorne 

Yeah, this was great.

Hannah Brainard

Thanks so much for listening.