Ep. 7, S2 - AI for Accessibility: Global Accessibility Awareness Day

I think it's very important to remember when you are someone making decisions about accessibility and not experiencing life with a disability that you need that input in the room as you are trying to make progress in that direction of being more inclusive. So as a government communicator with accessibility and inclusion goals, whether it's just to be compliant with ADA and DOGA guidelines, or it's part of a larger initiative that you have that goes beyond the basic compliance requirements. It's just important to keep that in your checklist. 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. Welcome back to the Government Video Podcast, where we explore how local government and PEG stations use video, media, and technology to inform, engage, and serve their communities better. I'm Michelle Alimoradi, and I am your host this week. We are keeping our NAB theme for this episode, but on this day, which happens to be Global Accessibility Awareness Day. We are going to talk about some AI for accessibility topics that came up at NAB this year. I'm super excited that NAB was talking about this this year because AI artificial intelligence has been the buzzword of the last few years, but we haven't been hearing from the NAB organizers, these conversations about the accessibility tools that are now. Being made possible, being made more affordable through the use of AI. So this year we saw a lot of rapid development in the use of AI translations. We are also seeing things like speech synthesis start to pop up. I went to this session that was called AI's Potential to Revolutionize Accessibility that was moderated by Joe Devon, founder of Ally Audits and apparently co-founder of this Global Accessibility Awareness Day. Shout out, Joe. It was a, it was a great panel. I felt fortunate to be able to attend. They were also bringing up the topic of AI A SL, which is something I've heard about before, but this was like the deepest we've, I've heard anyone get into that conversation at NAB and it was with someone on the panel who is in the deaf community. So it was very interesting to also get that firsthand feedback about how someone from the deaf community feels about these developments. And actually the whole panel was filled with people living with disabilities, so that was great to see. And intersectionally, they also brought their perspectives on technological development on the build side. We had perspectives from, you know, storytelling in general usability of the technology that's informed by people living with disabilities and their real lived experiences. We also had some legal perspectives, so it was overall just a really great panel to start, and they were all talking about their hopes and fears for the future of AI with respect to these accessibility tools. So I'm going to play a few clips of memorable moments that stuck with me during the panel and then give you, you know, my own takeaways and also talk about, you know, the impact that folks in local government communication can have on the future of these tools. Let us not forget our influence in this space. The first panelist that we're going to hear from is Storm Smith, who is a storyteller and multihyphenate influencer. She has been steering creative direction and offering her expertise on brand projects, collaborating with industry giants such as Amazon Studios, apple, Netflix, meta and more. And in this bit that I'm going to play here, we're getting into, of course, Storm's, hopes and fears for the future of AI. And I just think she brings up some really great points and has some really quotable moments here. So before I play this clip, know that Storm is a member of the deaf community and is signing a response to an interpreter who is in the room, that interpreter is speaking the response out loud to the people in the room. So that's whose voice you are actually hearing and a additional. Interpreter is then signing that response that the other interpreter is speaking for the people in the room. So we've already got a great accessibility process going on in the room for this session. So let's hear from Storm. Okay. So fears not so much to be honest, because I think human nature we always, and we just don't know what we don't know. Right. So that triggers that psychology in our body. We're psychologically, but we're already learning A, what AI can do and what it can bring. The benefits of work efficiency, saving time. So I see a lot of potential there, but we're, we still have a long way to go. And that brings it to my next point about the hopes. I always believe in the philosophy of you. Any work we do, assuming everybody here on this does that, but nothing about us without us. So really seeing the hope that they include people with intersectional identities, culture, intelligence, lived experiences that can contribute to the AI platform to make it improve and experiences from firsthand is what you need. Hire us to consult from the beginning to the end, have us in the room, and that's really where you'll see the true change. So I love what Storm says here, particularly this takeaway quote, nothing about us without us, because I think it's very important to remember when you are someone making decisions about accessibility and not experiencing life with a disability that you need that input in the room as you are trying to make progress in that direction of being more inclusive. So it's just really great to have a reminder about this. It's something that can be overlooked, not intentionally, but folks are worried about deadlines. They're evaluating a lot of tools and thinking about budgets, and it's easy to, you know, let something like this fall through the cracks. So as a government communicator with accessibility and inclusion goals, whether it's just to be compliant with ADA and DOGA guidelines, or it's part of a larger initiative that you have that goes beyond the basic compliance requirements. It's just important to keep that in your checklist. Think about it, when you are coming up with any new process in your department, either internally or in external campaign, think about it when you're evaluating vendors, because we can put the pressure on people who are developing these tools to ensure that the tools that are being created really are the best tools because we have very, very powerful technology to work with, and it can do a lot as long as we're asking the right questions and solving the right problems. So I think that was just a really great thing to point out, just to make sure that it stays on people's radar. And next we heard from our legal panelist, Ariel Simms, who is the president and CEO at Disability Belongs. And Ariel was getting into a more ethical part of the conversation. Let's hear what they have to say. Yeah. I'm a lawyer, so I think we're naturally afraid and, and risk averse and wonder about the ramifications of things. I think my fear is that AI will continue to grow and expand exponentially and not be sufficiently regulated that companies will not do what they need to do to work alongside and give the disability community on tools that are meant to foster accessibility. But on the flip side, I do have a lot of hope that AI can really close the gaps in accessibility that we have now, and that the number one thing that we hear from companies and organizations when we're working with them to be more disability inclusive and more accessible, the thing we hear is it costs too much. And that becomes a huge barrier for people to really start to think about accessibility and best practices. So I do have hope that as we continue to perfect AI tools and we figure out the systems for sufficient human oversight. That we can develop AI in an ethical way and that we can lower the barrier to entry to truly accessible and inclusive content. So I love what Ariel brings up here because again, these questions of ethics are something we all know, right? This is something that's been coming up in the AI conversation ever since AI went mainstream, that it's developing so rapidly that regulatory bodies cannot keep up, and they also can't anticipate what they are going to need to regulate in the future. So it's this wild west situation, but we do have influence. We do have the power to demand accountability. To demand statements about data usage, to demand transparency. We can demand that. We can influence this space in that way, and it can be easy to overlook that step if you don't currently have anyone living with disabilities on your staff or in an executive role. It's even more important to take note of that accountability step and put it somewhere in your process as you're moving forward so you can try to help with this ethics problem. And to speak to what Ariel said about cost, I think that's one of the things that makes this more difficult. People are so excited about being able to offer something that's now within the budget. Previously was very cost prohibitive that were not necessarily being as critical of the tools as we should be that we're getting in this place where we're like, oh, this is good enough. So we're just gonna apply this bandaid and then hope that it gets better. So our next panelist was Dan Caddigan, who's the Chief Technology Officer at Three Play Media and Dan is globally recognized for his contributions to AI enabled captioning and audio descriptions that meet rigorous accessibility standards and cares about these accessibility standards in his work developing these tools. Let's hear what Dan had to say when asked about his hopes and fears for the future, creating a lot of new capabilities that I can use my team. I'll get into a full AI audio description in a bit, but that's something that we seeing like science fiction even a year ago. So these things are coming on fast and they're really reducing costs, making it so that you can hit on backlog content. Really hopeful. As these things come along, there's a lot of folks like myself and my team that are ready to catch them and run with it. On the fearful side, I won't get started on a GI. That could be a whole other talk here, but from an accessibility perspective. I am fearful that quality will get good enough. Kind of like we're what we're seeing with a SR right now. And once it gets good enough, companies will go for that. 'cause the cost is one 10th, one 20th, one 30th of what you pay for. Can you off for a second? Can you define a SR Oh, sorry. A SR, automated speech recognition. So speech to text. The quality has gotten quite good. It's still not as good as human, not by a long shot, but it's gotten good enough where companies can say, it's really good. We're just gonna use that. Take the cost savings and move on. So if other AI technologies get good enough, I could see you make. So what Dan brings up this concept of being good enough is important, and it is the recurring theme that we're hearing, right? So from the start Storm is addressing it from, you know, if you're not consulting with us, you're gonna end up with these tools that maybe at face value look like they're helping, but. You didn't really talk to the people using them, so they're not as great as you think. And then with Ariel, we're hearing, you know, that same theme coming through, mentioning the cost savings. And Dan gets into that a little bit too, where people are so excited that something is now. Within their budget. It's not cost prohibitive anymore, and they're not evaluating it. They're just slapping it on and calling it good. There is a danger that if we do that and we're all just excited that, hey, we made the thing and people are buying it, that the tools will not ever meet their full potential to be useful and to actually promote inclusion in addition to having people with real lived experiences in the room, not just for people living with disabilities, but also the cultural context. Dealing with AI translation, we end up making a tool that in some cases is just learning how to be worse and worse because we're not. Being super intentional in the way that we refine it with time. It's important to be critical. Like so many of us are so impressed already with AI because it is really impressive. It is really cool. It's already saving people. So much time saving people, so much money opening so many doors for people in the accessibility space. As is, but it can be better, and we need to remain vigilant. We need to remain critical, and we need to find those ways where we can pull the levers as people who are not developers, as people who are not lawyers to make sure that these tools go in the right direction, because we all know AI could go in very bad directions. We've seen some of the examples at how fast it learns. Not very helpful things, so. Just keeping all of these things in mind as you build your process in your communications departments is very important, and that is what Global Accessibility Awareness Day is all about. So I think that this will be our last episode about NAB. So hold tight. We're gonna have some fresh new topics coming up in our next few episodes. Thank you so much for joining us to cover the full range of the benefits of NAB, the content from NAB, and thank you so much for joining me for this conversation about AI for accessibility. If you found this podcast helpful, I hope that you'll share it with a colleague. I hope that you'll like and subscribe on the platforms where you listen, and thanks so much. We'll see you on the next one.