Ep. 11 - Practical AI Tips for Local Government Communications Teams

Today, we are finally doing it. We are going to talk about AI. Are you ready? I'm ready. The discussion around AI and the role of local government communicators, specifically, is actually quite unique from other discussions of AI. Would you agree, Daniell? Yeah. I'd say our, motivations and our interests are different and our, reasons to maybe need to become knowledgeable and engage on it are a little different than what, you know, the business world or education or different groups. This podcast is for city communications teams and video professionals in government. We talk about expanding service delivery with video and streaming, media accessibility, gear, broadcast and streaming workflows, and more. It's all right here on the Government Video Podcast. Hello and welcome back to the Government Video Podcast. I'm Michelle Alimoradi and I'm your host this week. I'm here again with Daniell Krawczyk, founder and CEO of Municipal Captioning. Daniell works with organizations across the country on their video distribution and content accessibility. And we're really glad to have you back on the podcast, Daniell. Thanks, Michelle. I'm happy to be here. Always happy to have you. Today, we are finally doing it. We are going to talk about AI. Are you ready? I'm ready. And not just because all the cool kids are doing it either, uh, but because the discussion around AI and the role of local government communicators, specifically, is actually quite unique from other discussions of AI. Would you agree, Daniell? Yeah. I'd say our, our, our motivations and our interests are different and our, our reasons to maybe need to become knowledgeable and engage on it are a little different than what, you know, the business world or education or different groups. Yeah, exactly. Um, you can, you can Google AI right now and you're going to get, you know, infinite amount of search results. Um, but what you're hearing about in the discussions of other industries is completely different. Um, and we just really want to make sure that we touch on how it's important, uh, for local government communicators to be educated on AI, to keep residents safe, to keep public information safe. Um, but we don't, we won't be all doom and gloom today. We will also. Yeah. We'll be positive too? Yeah, we'll, we'll talk about some ways that AI could be, can be useful and save some time. But, we gotta get into this serious stuff first. Um, so tell us what's up. Tell us what's happening. Okay. Um, yeah, how to start? We were doing this install in Staten Island, uh, for Cablecast, uh, set up a year ago and learned, uh, an anecdote about the deer, um, swimming from New Jersey to Staten Island because they had run out of trees or, or, or, um, you know, things to eat. And that always comes to my mind over the last year of where, uh, money was suddenly drying up in the other areas where drifters were, uh, feeding. In the world of crypto cons and different things and then suddenly all the deer swam to the island of AI like a year, year ago, year and a half ago, something like that. So I think the first thing to start with is that beyond the fact that there's like crazy things happening with the technology and crazy things that con people could do using the technology. There's also an entire world of grifters and BS artists suddenly become AI experts or have started, uh, companies emailing you and telling you how they're going to use AI to do this or that for your company or your constituency. So I think the first thing I like to start with is that, um, having a good BS detector and cranking it to the most sensitive, uh, setting that it has, and, uh, assuming that, uh, you're going to be weeding through, uh, a lot of drifters and cons and hype, uh, when you're entering into anything around AI. Is, is, is wise because it's very important. Yeah. Very important. It's so realistic. Um, and a lot of people are combing through a lot of information coming at them and it's easy to get, um, you know, tired or less discerning and it's so all new and overwhelming that it's easy for someone to tell you that they know all about it and they know exactly what you should do and you're going to be able to save money here or make money there or do product here or there. And so I think the first thing I, I just need to communicate before we ever get into any of the ways this stuff becomes practical is that if someone comes to you and is telling you they can do something magical with AI or you're getting email about it or you're talking to somebody about it, just lead with your skepticism, double check everything, find somebody else who knows a bit more and have them run through stuff too, because it is a wide world of, uh, snake oil sales, masquerading as wise investment for major corporations, and I think that's the other thing. I will go ahead. Sorry. Oh, no I I'm just I'm nodding along because I agree and I'm sure there's probably a handful of people listening to this podcast that have already experienced this in some way. And the fact of the matter is that you can try to Google but AI is constantly changing. So it's not something that you can necessarily Google verbatim and get a direct answer about. So your advice about having conversations with other people who might know, uh, is extremely important for what you're talking about. Yes, exactly. Just, um, imagine it's like, uh, there's suddenly a bunch of new, uh, you know, get, get healthy quick, get rich quick things all over at once. Uh, and that's the world we're in right now with A. I. And I'm not even just talking about like people using A. I. To make images or write the emails or anything. I mean, 90 percent of the promises that are being made around A. I. That are getting all the major corporations spending billions of dollars on resources and investments are B. S. And there's going to be a bubble and it's going to pop the same way like that's dot com early web stuff, but will because the expectation around it and what it can actually do to help companies and governments make major, major changes to their operations and save a bunch of money is not panning out. I mean, this is right now we're in the, in July and, you know, financial magazines are pointing out that, you know, the 200 million bet becomes a 600 million bet looking at how much they've companies have put in. So I'm just pointing out right now. All of this stuff is, is ground shifting change that we need to pay attention and know about. And there's also going to be some big financial bubble where all the companies realize they wasted a bunch of money on things they didn't need. And that's not going to change the fact that we still need to be aware. We need to be cautious. We need to prepare our, our communities for the things they have to be careful of in terms of, of scams and things that take advantage of them. But separate than that, we're going to watch. The same thing that happened with like early web domains and things where there was a big financial bubble as the expectations didn't meet reality. We're gonna see that with, you know, AI. And we're gonna think like, oh, maybe we don't have to pay attention to it now that, uh, the financial magazines say it's not such a big deal. Yeah, it's It's always hard to discern when is the right time to get in here. Like, Oh, this tool has reached a point where it is actually useful and practical and not too unpredictable. Um, and I don't think that we've reached that point yet. And, um, I think what you're implying here is that public information officers, um, any local government communicators need to treat. AI with the same concern and care as other, like, PSAs, protective campaigns that they put out for residents to keep people safe, but it's, but it's not exactly the same, right? Because it's a moving target in a way that other scams have not been such fast moving targets. That's a great way to put it. Yeah. So do you do you have any advice? Yeah, how they might you know work with this ever changing I hate to say ever changing threat that It makes it sound insurmountable and I don't we don't want to be on here saying like oh, there's nothing There's nothing you can do like just be scared all the time. That's not the message that we're trying to send but it's vigilance is important and I think some people think that it's, if you don't engage with it, it won't affect you. But that's definitely not the case. Yeah. So I think, I think the first thing you brought up is like in local government communications, you've got a lot of roles. And one of those roles is helping to try to protect your citizenry, citizenry from threats. So when there's COVID 19 and you need to make a PSA about what's going on in the community around That or when there's, uh, you know, uh, flooding or anything like that. And then if, if there's things that we've noticed in our community, which I, I, I do our banking through a credit union. And when you go into the credit union, they have a big banner trying to keep people from falling prey to scams where they're wiring money or they're cashing a check, that's not really gonna clear. There's these things that, um, have been standard ways that people have been taking advantage of people over scams and phone calls and banking scams. And the first level we're gonna be dealing with and can easily be dealt with is the way that that's just expended one level higher now with AI. Now, it's a phone call that sounds like it's your grandson calling and saying they've been kidnapped and they need money wired somewhere as opposed to someone acting as if they're your grandson. They've, they've cloned the voice. They've taken the imagery, um, or it's faked video. And again, I'm talking about this right now in that like, you know, somebody getting tripped in your community into wiring 10, 000 kind of thing. But the same need to give people an awareness that they have to, you know, Be discerning and not fall for scams. Also is for like, if somebody is to make a fake video with the mayor making an announcement or, or a fake press release that's going out. So it's, it's, we have to actually start to prepare our communities to be discerning and double check the sources of things. Uh, if you see something that seems outlandish, don't say something. Talk to somebody else before you start saying something and telling everybody how you saw the mayor shutting down the main road for pedestrians or whatever. You know, uh, We can start with that because that gives us a way to talk to our community about the fact that this is as big of a ground shift as the internet. The internet changed things in ways we couldn't really tell what was going to happen, but we knew it was starting to change. Exactly. I think that's an important point to make and, and I think it's important to emphasize, I don't know the answer to this, but to emphasize that ultimately what we need to find, to as we try to navigate and educate people about possible scams and bad actors is that there's this level of balance that we have to try to maintain right? We have to find a way to not be alarmed immediately by the idea of someone calling and sounding like a loved one that needs help that's not real. We have to be able to train ourselves to have that moment. Um to react and and so as a as a Communications officer, how are you going to provide that kind of messaging? And then also thinking about, yeah, how are we going to get to the point where people are taking a minute to think about the fact that something could be spoofed and not get to the point where they start thinking everything is spoofed. You know, we've seen these issues like during COVID 19 where people just started to not trust facts and AI really kind of accelerates that. Um, that possibility of people getting like really jaded and not thinking that they can trust anything. So that will be the challenge, I think, for public information officers is, is find, striking that balance between how do we educate people without making them think that everything is fake. Um, and, um, That's an interesting challenge that we have not had to face in the past, uh, in a way that is very, very realistic, um, really plays with, you know, With the ways that our brains are, um, trained to react. Hi, Michelle here. We hope you're enjoying this episode of the Government Video Podcast I'm taking a short break from today's discussion to remind you that this podcast is brought to you by Cablecast Community Media. We're a cross platform video solution from Tightrope Media Systems. And despite our name, we help cities, towns, and other local media organizations get their video to viewers on all digital platforms, including cable. If your organization operates a cable channel, um, but we help you reach residents whenever and wherever they watch. Be it on your city's website, on mobile streaming apps, on over the top platforms like Fire TV, Roku, Apple TV, or even on their mobile phone. And we even help you push your content live to social media. We've been helping small non commercial stations launch professional, affordable, and efficient video initiatives for over 25 years. And our customers love us because we have great customer support and we have fierce industry loyalty. So if your organization is in need of cross platform video automation and delivery for local coverage of city council meetings or other local events, reach out to our team and schedule a demo today. And if you do, be sure to mention that you heard about it here on the Government Video Podcast to receive a special discount. That's all for me for right now. Back to the episode. Yeah. So. And I think on top of that, um, I think we maybe have a role in, in being really clear when the things that we are displaying are made with AI. You know, if you use AI, uh, to make graphics that you put up on your, on your bulletin board or that you accompany a blog post with, um, you know, saying this created with AI, because you're wanting to try to educate people to be able to decipher that, um, I, I've got a lot of connections on Facebook that are not quite so discerning, and I see people sharing very clearly AI generated images of sandcastles and things, and then saying, wow, look how amazing, how amazing this is. And I think, okay. Yeah, the Facebook, the Facebook shares, yes, um, I know what you're saying. And right now it seems innocuous, because it's just like a giant sculpture that couldn't have been possibly made, but it's, uh, getting people to be, um. Discerning and understanding. Um. Yeah, because that could just as easily be an image of a protest that didn't actually happen or, or something even more serious. Yeah. Um, one thing I do, we were talking about not going only down the doom and gloom, and I did want to kind of talk about where I think this stuff can be useful or helpful without, without dipping your toe into, Oh no, it made up facts. Oh no, I asked it for this and it gave me information that's totally incorrect. Um. And, I find for myself, I, I tend to write things in a less than clear way. I tend to explain things in three sentences that should be in one sentence. And maybe, and maybe backwards. I love writing to express my thoughts and get them out of my head. But then I think the end point of a lot of the writing is also supposed to be that other people read it and understand it. End. So it would be the ideal situation. Yeah, ideally. Ideally. I don't always lean into that, but I'm trying. So what I have found is, um, properly applied, you can ask the AI to read what you write and tell you what here is most unclear. How would you make some of this more clear? And then instead of just copying what it said, I look at it and go, oh yeah, you're right. I should structure the sentence with the Subject here because this makes it half as long and very clear then I rewrite it myself in a clearer way Something that you as a human like already have the skills to say There's no fact checking involved right like you as a human who have been a verbal communicator your whole life Can look at that and say hey like thanks AI for for helping me Do something that might have taken me an hour to figure out before in five minutes. Um, and I think that's really important to make those distinctions. Yeah, and then I think the hard part is, I think we're, as, as a species, as humans, we're then very tempted to select, all, copy, paste. If you, if you see, they have searched all the academic papers in the last couple years, and there's so many that say, as a large language manual, uh, I'm not, uh, uh, As a language model, I'm not able to help you with blah blah blah because people didn't even remove that part. So, it's very tempting to take what you get back and say, alright, this is good. But you, you can't, you, you have to yourself re write it. You have to yourself read it and, and, and type it all out, I think, because otherwise you're just very quickly gonna miss the fact that it actually inserted something completely wrong. Like, it's, there's parts that it's helpful, and then you have to be, or redundant, or, yeah. You can see that, yeah. So, and it's not like every word that is, you know, sometimes it's like, oh, this would be better this way. And you're like, no, that doesn't make, that's not accurate. But again, that's, that's an area where I have found it really helpful for myself is to, because all it's really good at is helping you conform to the expected, when you're not so great at conforming to the expected with your communication, it can help, uh, help funnel that closer to what everybody's expecting. And I think for PIOs and communication situations, that can be really helpful if you're trying to say, how can I make this more clear? Definitely, because you might have to bring in, you know, additional staff or other tools normally to do something like that, whereas you've got this tool now that can pull from all these resources to help you, um, you know, to give you suggestions on the way to make a very important message very clear first time around. And I think it can also be really useful in situations like just battling the blank page, right? Like you need to write something, um, and you can use it to just, without editing yourself at all, um, throw in, you know, how do I. Like, write me an outline for this speech that I need to make or something. You know, and give yourself a place to start. Um. just thinking, it's like a metronome for, for the writing. It gives me something to write to if I'm like, alright, I'm trying to accomplish this. Help me with a, a, a, an outline. And again, I, I do a lot of writing now for things that are like RFP responses. It's not creative writing. I'm trying to give them back something that addresses the things that they asked us to address. So, being able to say, what's, yes, where's the, give me the outline so that I know what structure I am. Otherwise, I might spend ten pages on the first part and never mention the second part. Again, I think what we, uh, what we see when we look at the world of art with AI, is that there's a small number of people who are engaging with it as artists, back and forth with a lot of intention, really making something very unique. Uh. Dr. Natrice Gaskins, who spoke at the Alliance for Community Media, still, you know, leading the pack at something that other people could try to copy, but you don't really even see anything close to that stylistically. But then you see other things that are just like, I clicked it, I took the thing right out of it. Uh, I think there's also a huge, huge difference in like using, uh, the GPT to kind of help you polish your government communications and, and anyway, just trying to rely on it, doing it for you. You know, there's no. There's no replacing anybody's job with this. There's only brainstorming a new layouts for your graphic. Sound the air so that you can then design it in a different way and you know things that are that's yeah I think that's a very important distinction is we're using it to save time in that sort of polishing phase right using as a save time in the polishing phase or or we're using it to save time in the initial brainstorming phase and Every step in between, we're still using our human brain and discernment and, um, and knowledge to make sure that everything that's being disseminated is accurate and not problematic and not confusing to people. I'll say this one thing, though, about the problematic thing. Uh, I, uh, Sometimes really good at seeing the things I want to see and not trying not catching where other people would find an issue with it So it's another area to just say hey you can feed it and say What, what are the three things, or give me some things that, you know, someone might find issue with, with this, or What are the most likely things people are going to complain about with this, you know? Even though if you're not going to address all, just to give you a bystander, like, okay, someone might complain that this language, I missed being inclusive here, or I missed referring to the Yeah, the mayor's office definitely could use tools like that, for sure. Yeah, I mean, again, it's, it's, I think it can be useful at, at getting ideas for what you wouldn't normally see, even if you're not going to, um, give it the keyboard by any means. Right. And, and the conversation doesn't have to stop there, right? You don't have to just have this conversation with this, uh, language modeling tool, right? You can, you can take those suggestions and then go talk to other people in your department about whether or not they agree. And it's just a way to get some. seeds, uh, for the ultimate. You know, end result that you need. Um, the last thing I'd want to say is I got to give a little presentation at the Wisconsin Community Media Conference about, about AI. And one part we talked about there that might be slightly relevant here for the government video podcast is that, uh, the best example I've seen of, uh, Community station experimenting and experimenting in a lot of different ways is the radio station WFMU out of Jersey City because it's a 24 7 broadcast radio station and each of the DJs have, if they wish, chosen to experiment in different ways and then it gets kind of riffed on in other people's programming. So. Someone, uh, one of the DJs, Michael Shelley, generated a bunch of AI jingles for his show. And then on the show that's called Tectonic, where they talk about Big Tech and the threat to the world of AI, they actually dug in and discussed, you know, is this bad? What is bad about it? Is it, what, what, it's actually, you know, attempting to engage in it as a discussion, which I think is what we really have the best opportunity for. What you and I are doing right here gets replicated by the government video folks when they're doing their own local discussion shows. And they could probably have, I can guarantee they could have a better conversation by grabbing a couple people from their community and asking them what they want and what they don't want, what they like and don't like about what's happening than, than any other subject they could, you know, pop, vox, vox pop a couple people out of the community. So, I, I really think, uh, trying to replicate the way WFMU is, uh, Just experimenting, you know, uh, uh, one of their shows is kind of a stunt radio show, uh, and they had callers have to guess whether the punch lines were generated by professional comedian or the A. I. And I just, I think you could do so much with, uh, with your own local community by just, You know, turning it into a little bit of a game show, turning it into an opinion show, turning it into a cross talk. We have two people that are vehemently opposed and two people have been playing around with it. Absolutely love it. And we're going to give them an hour to discuss it here in the middle of our government studio. What I what I love about that example is that it really goes back to this sort of scam prevention and discernment that we talked about at the beginning, right? So it can be in, in good fun, um, and keep things lighthearted, but at the same time. You're also giving people the experience of having to think about why they think one is real and one is not and having, like we said, having that moment where you just take a second to be like, wait, don't take this at face value. Take a second to, to think about it. Like check in with your gut on this, you know, um, because we are, I think Call a friend or family member rather than respond, right? Yeah, I think we're still humans. And I think we can still recognize the difference between what we do and what these language models do or what these voice, you know, spoofing programs do. I think that we can still do it if we just practice and flex that muscle on a regular basis. So I really love that example for that reason. And I, and I hope that our folks in, in local government are hearing this and thinking about ways that they can educate their community without. Terrifying everybody, because I think that's an important, uh, aspect of it. Yeah, trying to do it in a way that would be fun. I like, I like how you put that. Uh, it got me thinking of a Game show that you would have to guess which of these kidnapping audio calls are real. Is this actually your grandson or not, Thelma? Maybe that's not a great game show title. Maybe not that one exactly, but I like that direction you're going with it. Well, Daniel, it's always a pleasure having you on. Oh, thank you. Appreciate it. This is such an important topic and I'm sure we will touch on so many more facets of it. of AI in the future. So, uh, hope you'll come back and talk to us about it again. And, um, we'll, we'll talk to you again on the next one. Awesome. Looking forward to it. Thank you. Thanks so much, Daniel.