---
id: "workshop:using-artificial-intelligence-to-de-jargon-government-language"
type: "InnovateUS workshop"
title: "Using Artificial Intelligence to De-Jargon Government Language"
url: "https://innovate-us.org/using-artificial-intelligence-to-de-jargon-government-language"
date: "2024-03-07"
duration: "60 minutes"
language: "English"
instructors: ["Abigail Fisher"]
recording: "https://www.youtube.com/watch?v=2Lk0MTXzbck"
transcript: "automatic, from the recording"
source: "InnovateUS (innovate-us.org)"
licence: "CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/), as stated on innovate-us.org"
---

# Using Artificial Intelligence to De-Jargon Government Language

InnovateUS workshop · 7 March 2024 · 60 minutes

Join us for this 60-minute virtual workshop with Abigail Fisher, civic design expert working with the Colorado Digital Service, and former design lead at New York City's Civic Service Design Studio. In this training focused on making government communications simple, comprehensive and effective, participants will learn about the importance of using plain language for communicating with residents and in which context and circumstances using plain language is necessary and appropriate. Participants will also learn how to use artificial intelligence (AI) tools to de-jargonize government language, with a hands-on tutorial on how to turn a complex piece of government writing into simpler and easier-to-understand language for government employees and residents alike.

## Learning goals

- Learn about the importance of using plain language for communicating with residents
- Learn how to use artificial intelligence (AI) tools to de-jargonize government language, with a hands-on tutorial on how to turn a complex piece of government writing into simpler and easier-to-understand language for government employees and residents alike

## Speakers

- **Abigail Fisher**, Civic Design Expert, Colorado Digital Service

Recording: https://www.youtube.com/watch?v=2Lk0MTXzbck

## Transcript

Automatic transcript of the recording: names and terms can be misheard, and the times can be a few seconds off. Each time links to that moment of the video. The transcript alone: https://innovate-us.org/using-artificial-intelligence-to-de-jargon-government-language.txt (plain text) and https://innovate-us.org/using-artificial-intelligence-to-de-jargon-government-language.json (every segment with its start and end second).

**Unknown speaker**

[0:01](https://youtu.be/2Lk0MTXzbck?t=1) Spotlight you. Yeah.

**Abigail Fisher**

[0:05](https://youtu.be/2Lk0MTXzbck?t=5) So great to be here, everyone. I'm so excited to be in community with everybody today. And Hannah, is my slide deck looking okay?

**Unknown speaker**

[0:16](https://youtu.be/2Lk0MTXzbck?t=16) Yes. Yes, it looks great. Fabulous. All right.

**Abigail Fisher**

[0:22](https://youtu.be/2Lk0MTXzbck?t=22) All right. Hi, everyone. So, like Hannah said, my name is Abigail Fisher. I use she, her pronouns. And I'm so excited to be here today to talk to you about using artificial intelligence to de-jargon government language. And I want to pause and talk about this image that I created for this slide deck. I actually created using MidJourney, which is an image generation AI tool. And the prompt that I used was, please show me a community co-designing a garden using AI in partnership. And this is what it gave back to me. And I really think it's a beautiful image. And I think that that idea of sort of AI tooling as partners to us is very much a thread that I hope you kind of see throughout the day. And hopefully this is kind of a future that we might be building with these types of tools in community with one another. So, like Hannah said, my name is Abigail Fisher. I use she, her pronouns.

[1:22](https://youtu.be/2Lk0MTXzbck?t=82) I am a queer civic designer, and I live with obsessive compulsive disorder. That lived experience really comes into play in my work with Colorado's Behavioral Health Administration, where I kind of marry that lived experience with my design skills to really encourage and co-create behavioral health technologies with the folks that are closest to those issues and services in Colorado. And so in this scenario in which I'm currently working in, the way that I define community are the behavioral health providers that give care. And then the communities that are looking to receive or engage in behavioral health care. So mental health services and substance use services. Yeah, so I've been across a couple of different states. And another thing I would really love to hear from folks in the chat is like, have you used AI tools before, plain language tools? And for what types of projects? So if you're feeling called to that, I would love to hear a little bit of that discussion in the chat.

[2:28](https://youtu.be/2Lk0MTXzbck?t=148) What tools have you used? Have you liked them? And kind of how have you been using them to date? Or not, maybe you have it, and that's totally okay too. So we have a couple of goals on our agenda. And it's really important for us to both define shared language together, because without that foundation of shared language, we can't move to broader understanding or action. So we're going to begin by defining and then encouraging use of the plain language as a framework. We're going to hold space so that we can practice with AI and plain language tooling. Because one of the things that I think is really important, especially being in government work, because our work is such high priority and high stress, there's never room for us to practice. So that's actually something we're talking a lot about in Colorado is how do we hold these spaces to practice? And I love this workshop series because I really think it hits on that value.

[3:28](https://youtu.be/2Lk0MTXzbck?t=208) And then finally, we're going to discuss some government specific watchouts. So those that work in and around government will have different sort of criteria and parameters in which we can use AI tooling. So I'm just going to give a couple of cautionary call outs at the end of the presentation. And then, you know, this is such a caveat that I'm sure I don't need to use, but, you know, trying not to fall into that imposter syndrome. But my expertise really lies in design, community building and technology. I don't consider myself to be an expert in AI or plain language necessarily, but I do use both of those things in my work frequently in design and community building. So just wanted to give kind of that background on where my expertise falls and, you know, why I have the authority to talk about certain things and will point you to resources for other topics. All right. Oh, I'm seeing these chats come in. That's so great.

[4:29](https://youtu.be/2Lk0MTXzbck?t=269) Sky says that she's used these tools to help get ideas on how to make things more plain language from documents to presentations. That's incredible. That's what we'll be talking about today. David says, Chat GPT, Bing, Gemini. So this is great. Thanks for sharing. And I love, Sonia, the research travel plans. I do that, too. I just planned my trip to the Code for America summit using Chat GPT. So it's awesome to hear the wide ranging ways that we're already using this tooling. But to begin, I just want to talk a little bit about the fact that what we do is so complicated. Even going through my resume, when I try to tell people like what I do, people don't understand. And that's even me kind of like in a bar trying to communicate. But, you know, at the end of the day, AI tools that marry AI and plain language can help us to keep that jargon and bias out of our language so we can communicate more effectively with our communities.

[5:32](https://youtu.be/2Lk0MTXzbck?t=332) I know that in my job, I'm often handed down these hundreds and hundreds of pages of legislation. And that legislation has direct impacts on how people receive behavioral health care to effectively engage with the communities and bring them into the co-designing process of those technologies. The first step is to sort of de-jargon that legislation and say it in a way that communities immediately understand and they understand how to engage with that particular bill, policy or program. And like I said, you know, plain language is a great framework to help us to get there. So this is a great definition from plainlanguage.gov. Essentially, plain language is a, you know, a framework or techniques to ensure that your audience understands something the first time that they read it. And you can read a lot of great techniques that people use on this slide, active voice, short sentences.

[6:32](https://youtu.be/2Lk0MTXzbck?t=392) And I'm sure that many of the folks in this room that are familiar with plain language remember how this used to be. You sort of had to embody these rules in your writing. You had to memorize them and practice. And, you know, now we are seeing, you know, in the last five years, we've had a lot of great tools that have helped us to score writing for plain language. We'll look at those in a little bit as well. Some of those are the Hemingway app or TextBuddy. And they give you a score and they tell you this is how close your copy is to being plain language. But now with AI, the cool thing is we can do that and go one step further and partner with AI as a writing partner. And again, that writing aspect is sort of the new aspect of AI tooling that we'll be talking about today. But yeah, in short, plain language is just a powerful framework to help us check against jargon, acronym, and legalese.

[7:32](https://youtu.be/2Lk0MTXzbck?t=452) And should really be our government organization standard for communication and outreach, both internally and externally. And this is just a fun question I love to ask, especially to government groups. If someone has a really long or funny acronym, I would love to hear it in the chat. Because I'm sure we all have long and funny acronyms that are definitely not plain language. But are things that us in government have to deal with frequently. So any funny acronyms I would love to hear. And yeah, so I also wanted to take a moment to kind of define and talk about what do I mean by communication? And to ask the question, how do you communicate and to what audiences? And I did a brainstorm of some of the ways that I communicate on a daily basis. Everything from emails, to official legislation and bills, to printed pamphlets, to surveys, to social media.

[8:33](https://youtu.be/2Lk0MTXzbck?t=513) Those are all ways that we engage with our various communities that are important to us. Whether those are policymakers, legislators, community-based nonprofits, or activist groups that we're really trying to engage in and making change with us. I'm curious, and people can come off chat, I would love for this to be super free-flowing, but did I miss anything here? How else do you communicate with your communities? And what audiences are you speaking to most frequently?

**Audience**

[9:08](https://youtu.be/2Lk0MTXzbck?t=548) Maybe I missed it, but we use a lot of maps to communicate.

**Abigail Fisher**

[9:12](https://youtu.be/2Lk0MTXzbck?t=552) Maps, oh, that's such a good one. Who shared that? I didn't see who came off of mute, but.

**Audience**

[9:20](https://youtu.be/2Lk0MTXzbck?t=560) Sorry, this is Matia Yepsen from New Jersey.

**Abigail Fisher**

[9:23](https://youtu.be/2Lk0MTXzbck?t=563) Thank you so much. Maps, oh, that is such a great one. Anything else? I'm looking, focus groups, community meetings. That is another good one. Meetings. This is awesome. I'm looking at these acronyms too, and Sarah, that acronym is one of the funniest ones I've ever seen, so that's so awesome. All right. Presentations at forums, tabling events. Site visits. These are also great. So, yeah, I think these are really awesome. And I just want to say that, like, everything we've listed, including maps, including emails to your colleagues, those are all appropriate places to employ plain language. One of the things that I like to say is, like, plain language is an act of love. And, you know, using plain language internally, where your colleagues maybe don't understand your program or your policy to the same extent that you do, can be a real act of love in creating more understanding, even internally at your organization, even if maybe external communication isn't necessarily a part of your day-to-day responsibilities.

[10:42](https://youtu.be/2Lk0MTXzbck?t=642) All right. Well, this is an AI tooling workshop, so now we're going to make this turn to, like, how does plain language meet AI tooling? And, again, just to set some shared definitions of what is AI, I have a definition of both AI and machine learning up on the screen. I often find that folks in technology, in government, especially in high offices, like to try to, like, weaponize these technical terms and say, hey, this is too complicated for you to get, just understand that we get it. And I want to tell every single person on this call, you absolutely can understand AI, machine learning, and any other technical concept that comes across your desk, and there are places like the Journalist Organization, the Markup, that is helping to do just that. So just wanted to throw that out here. And just to quickly differentiate between AI and machine learning, AI, again, is really a system for a set of human-defined objectives that can make predictions and recommendations. And then machine learning is an aspect of AI that is the mechanism for making predictions and recommendations.

[11:46](https://youtu.be/2Lk0MTXzbck?t=706) So that's that, you know, differentiation, just in case anybody ever asks. But we're going to be looking at a number of AI tools that marry the principles of plain language with the principles of machine learning. We know that there are a couple of different types of AI tools, well, not a couple, there's many, but at broad strokes, there are a couple of different types of AI tools. And we're going to be looking at a number of different AI tools that can be used to make predictions and recommendations. And then machine learning is an aspect of AI that is the mechanism for learning and improving that AI system. So that's that, you know, differentiation, just in case anybody ever asks. But we're going to be looking at a number of different types of AI tools, well, not a couple, there's many, but at broad strokes, there are image-based AI tools like MidJourney that I use to kind of generate that image at the top of my slide deck. And then there are text-based generative AI tools like ChatGPT, where you're more in dialogue with the system. And I'll show you what that looks like in just a minute in a demo. But I want to say that there are, you know, thousands and thousands of AI tools on the market.

[12:49](https://youtu.be/2Lk0MTXzbck?t=769) I'm not here to sort of endorse any one given tool. But I do have a couple of suggestions of questions that you can ask as you engage with these tooling. And these are questions that I ask myself before downloading a new tool. So the first question, is it free? And if it's not free, why? What is their profit model? And how will that impact you as a user? It's kind of a similar question, is it open source? If it's not, why? I'm having some interrogation around that. I personally am really trying to move myself to only technology tools that are free, open source, and accessible. That's kind of like my personal code. Take it or leave it. Next question, how are they using my data? Look for a data use statement. AI tools, your data becomes part of that benefit of that AI tool. So it's really important to know how is your data that you're putting in being used? And then finally, what is the source material? And we'll talk about this a little bit when we're comparing different plain language tools today.

[13:53](https://youtu.be/2Lk0MTXzbck?t=833) But for instance, ChatGPT has a massive database that they're pulling from as you're engaging in conversations with that tool. And even when I asked ChatGPT, and we'll talk about this in a little bit, what source do you use to define your definition of plain language? ChatGPT can't answer that because it's such an aggregate, right? Which is interesting in one sense, but can be problematic in another because you don't really know where the source of that framework comes from. Now, there are other tools that are more bespoke and nuanced where you do get more understanding of like what their definition of plain language is. And what are they using as their source material or their database? And, you know, I think that Innovate US actually has a workshop on this where you can make your own AI tool where you can control for all of those questions. So that's just a prompt to say, like, be a little skeptical, ask some questions before jumping in. And again, we'll talk a little bit about AI ethics and government at the end of this call.

[14:57](https://youtu.be/2Lk0MTXzbck?t=897) All right, so we have one big... activity for the day. And this is kind of my cue into that activity. So the activity we're going to do, and I have an instruction slide in just a second, will be to go to your organization's About page, pull a paragraph of text, and then I'm going to ask you to use an AI writing partner to edit that paragraph of text to be more in plain language. And I actually did this with my own bio, which I fully admit is full of jargon and government specific language. Like I said before, when I try to describe what I do in a coffee shop, doesn't always go so well. So it was actually awesome. And I gave each of these tools, so TextBuddy, Hemingway Editor, and ChatGPT, essentially the same prompt. And you can kind of compare and contrast how these different tools changed up my bio. And I'd honestly love to hear in chat which of these you like most, because I'm not really sure yet either. But there's some interesting interpretations here.

[16:04](https://youtu.be/2Lk0MTXzbck?t=964) All right. But yes, so like I said, I'm going to play a short demo. I'm actually going to talk over this. So again, ChatGPT is a free tool that you can use. So if you haven't set up an account, ahead and do that. This is me logging in to my existing ChatGPT account. And if you're using ChatGPT, you can start out with a prompt asking them to rewrite a piece of text using plain language. You can see that I'm asking ChatGPT to help me rewrite my bio. And then I go ahead, and you'll copy that piece of text into the tool. And depending on what the tool is, you can have some back and forth dialogue. You can see that here, it spit out my rewritten paragraph. And then I go on to ask some questions about like, can you make it shorter? Can you tell me where the source of your definition of plain language came from? So this is what I'm asking it right now.

[17:07](https://youtu.be/2Lk0MTXzbck?t=1027) It's pretty interesting answers when we see those pop up on the screen as well. But again, maybe while you're watching this demo, go to the organization that you work at or work for, go to that About page, grab a paragraph of copy. And then again, you're going to use this as a writing tool and try to simplify that paragraph. So we'll do this for about five minutes. And then I'd love to come back and discuss as a group, how did it feel? What road bumps did you hit? And if you feel like the paragraph you end with is better than the paragraph you started with. Any questions about that activity before I let you off for five minutes? Yes. So the other, that's a great question. So let me go back to this slide. So these were three tools that I picked TextBuddy and Hemingway Editor. I chose because they have more of a plain language bent. So you can go and use for free TextBuddy and Hemingway Editor's plain language scorecard, but you have to pay for a premium for their AI writing assistant.

[18:26](https://youtu.be/2Lk0MTXzbck?t=1106) So up to you, which ones you want to play around with. I think they're all probably equally interesting. And of course you can use another one that I'm sure I haven't heard of as well. Alrighty. So I'm going to put on a timer for five minutes and this will ring when we hit that five minute mark. So I'd love to come back after five minutes and talk to everybody about how their edits went. Hannah, anything I missed?

**Unknown speaker**

[19:09](https://youtu.be/2Lk0MTXzbck?t=1149) Nope. Just wanted to apologize. I think some folks, we are, didn't receive the right Zoom link, but I've been sending them individually. So just wanted to say apologies for that inconvenience and I'm glad you're all able to join.

**Audience**

[20:21](https://youtu.be/2Lk0MTXzbck?t=1221) Could you paste the prompt in the chat?

**Abigail Fisher**

[20:26](https://youtu.be/2Lk0MTXzbck?t=1226) Yes, absolutely. Let me, I might have to.

**Unknown speaker**

[20:51](https://youtu.be/2Lk0MTXzbck?t=1251) And Abigail, if you could just repeat the prompt, there might be a few people who just joined. And again, very sorry about the Zoom link. But if you could just repeat the prompt for folks that are newly joining and thank you for all of your patience. Yeah, thanks everyone. Yeah,

**Abigail Fisher**

[21:28](https://youtu.be/2Lk0MTXzbck?t=1288) no worries at all. So the prompt is simply to go to whatever organization you work at, go to their about page, grab any medium paragraph of copy about, you know, let's say five to eight sentences and then paste that text in a plain language AI tool of your choice. We gave three recommendations, TextBuddy, Hemingway app, or ChatGPT. And then we're spending about a couple of minutes editing the text with the AI as a writing partner to make that language a bit more clear and direct. So just a quick, you know, editing exercise.

[23:02](https://youtu.be/2Lk0MTXzbck?t=1382) All right. About 30 seconds left and wherever you're at after that 30 seconds is perfect. Did everybody could everybody else hear that very calming noise? I like that. Okay, good. All right. So I'm going to stop screen sharing for actually a minute. And I'm wondering if anybody would like to come and share about how that felt for them. And what was the the prompt that I put into chat GPT and I'll actually link to it as well. I you can save out chat GPT dialogues so I can. It might have been that Stardew Valley. Now I'm getting distracted by the chat. So Kim, I'll put my chat GPT thing in before but just to kick us off like how did that feel for people? Did that feel? Did it feel good? Confusing? What were some of the results that you got? Did they make sense? Not so much. And folks can come off mute or go in chat.

[25:00](https://youtu.be/2Lk0MTXzbck?t=1500) Ervin, I see yours and then Nancy's hands raised. Ervin, do you want to go ahead?

**Audience**

[25:07](https://youtu.be/2Lk0MTXzbck?t=1507) Sure. So I put in our our business outreach wording trying to attract women's businesses, small businesses, minority owned businesses and things like that. The I used Copilot from Microsoft, and it reduced the content into plain English. That's what I asked it to do. But in the process, it left out some of the wording that we were very deliberate about trying to convey that we care about women's women's businesses, small business, and up minority business, it just compressed a lot of that out of the messaging. And it transformed it from short paragraphs to bullet bullets, bullets, but it left a lot of information out. So I can show you the screenshot if you want, but that's the basic thing that I discovered.

**Abigail Fisher**

[26:05](https://youtu.be/2Lk0MTXzbck?t=1565) Yeah, up to you. I think that's such a great point, Erwin. And I think it's really why we talk about AI being a writing partner versus someone who's taking over our writing, because a machine is never going to really understand, you know, why you emphasize certain words that you did. So there's always going to need to be that partnership. So I thank you for bringing that up. I think that's an excellent point.

**Audience**

[26:29](https://youtu.be/2Lk0MTXzbck?t=1589) Sure. Would you like to see the screenshot? Or you just have to share? I have to share my screen.

**Abigail Fisher**

[26:33](https://youtu.be/2Lk0MTXzbck?t=1593) Maybe for the sake we can just do audio if that's cool.

**Audience**

[26:37](https://youtu.be/2Lk0MTXzbck?t=1597) Okay, that's fine.

**Abigail Fisher**

[26:39](https://youtu.be/2Lk0MTXzbck?t=1599) Yeah. Nancy.

**Audience**

[26:43](https://youtu.be/2Lk0MTXzbck?t=1603) So I did the about page for Colorado's Department of Health Care Policy and Financing. And I thought it was interesting. I think it did make the language more accessible. We talk about mission and took that out and made that language more clear. But there was one thing that made me cringe was, and I think it's because I always got dang for it in college, was using the word yes. So I asked it to please rewrite it not using the word yet. And so it took the sentence, our goal is to make sure everyone gets their access to health care. We're changing to our goal is to make sure everyone has access to health care. And so I liked that it tweaked it to fit my own personal taste.

**Abigail Fisher**

[27:42](https://youtu.be/2Lk0MTXzbck?t=1662) Yeah, thanks, Nancy. That's super interesting. And I heard people talking in the chat. And I wanted to bring this up as well that Zach said, you can also tell these tools to keep certain pieces of language. And I think that that's a great point. So in the sort of demo that we gave, I only kind of gave it one shot, but you can really keep editing and editing again by asset make it and sometimes you can even say something like make it friendlier. Make it less scary. And it's interesting to see kind of how how the tool interprets that. But thanks for sharing, Nancy. That was awesome. Thanks, Nancy. I also see Jamie's hands raised.

**Audience**

[28:24](https://youtu.be/2Lk0MTXzbck?t=1704) Hi, I don't know if this was said already. But just something small, I noticed that when our bio is used, or like in the bio, if our company name was used, like repeatedly, in a paragraph, chat, gbt would remove it and replace it with things like we so I don't know if that was like, meant because of plain English accessibility, if that was just like something it did to be concise or like look better, but I noticed that.

**Abigail Fisher**

[28:50](https://youtu.be/2Lk0MTXzbck?t=1730) That's a great observation, Jamie. And yeah, again, that's the thing. I think like plain language is not a monolith, right? Like there are many different interpreters of plain language. And although there's kind of that, like key guidelines, how that looks in practice can be really different. So I think it's smart. Oh, I had the timer was on again. And I heard I don't know if everybody else could hear that. More Stardew Valley for folks that that like that, which is a video game that is very fun. Anybody in chat or like hands that I can't see maybe before we move on. This has been a great conversation, y'all.

**Unknown speaker**

[29:36](https://youtu.be/2Lk0MTXzbck?t=1776) So there's a couple of comments in the chat, no more hands raised. But some Yeah, I encourage folks to drop into the chat because because others are sharing how they what it changed for them and how they used it.

**Abigail Fisher**

[29:50](https://youtu.be/2Lk0MTXzbck?t=1790) I love it. Y'all are the best. I really I'm loving this dialogue today. It's been super fun.

**Audience**

[30:00](https://youtu.be/2Lk0MTXzbck?t=1800) Abigail, I would just make the point, you know, we have pretty good writers in our government agency and everything we do is tends to be crafted to touch different audiences we have and all of that and so trying to un-simplify that can also be a problem because if you say I want this audience and that audience it'll get mixed up and so I think good writing is actually good writing if

**Abigail Fisher**

[30:32](https://youtu.be/2Lk0MTXzbck?t=1832) you can do it and simplify it. Totally, what a great point. I think if we had like two, a second part to this series, that second part could be exactly that, like how do you differentiate tone between the different audiences you have to speak with and when do those kind of overlap nicely and when are those so dichotomous that you have to have different communication strategies and the one that I always think of that I experience the most in my job is how it's so different to talk to legislators than it is to talk to government staff and it is to talk to constituents, like those are, you know, there's overlap there but there's definitely a lot of different pressures and things that those folks care about that you have to communicate differently so that's a really excellent point. All right, well I have a couple of slides left and then I think we can keep going into more of a dialogue and conversation because I'm loving this vibe but these slides are honestly a lovely compliment to some of the things that we have been talking about. Although AI can be a really powerful tool for co-writing and thinking and having us write and talk to one another in different ways, in ways that maybe we don't necessarily embody just in our bodies but could partner with a technology system on but I think as we've all talked about today and many people have given examples, AI is not perfect.

[32:06](https://youtu.be/2Lk0MTXzbck?t=1926) AI tools can in fact cause harm unless we understand that these tools are not infallible but are really only as good as the source material and as many people have been saying in the chat, only kind of as good as like the prompts that you're able to engage the tool in and an example of this is that in another AI tool that I put my bio in, it called me using the phrase queer civic designer with obsessive-compulsive disorder as a negative phrase that I should remove which we all know that being queer and having OCD is not a negative thing and it's actually something that's really important to my bio and I think some folks in the chat shared similar examples where the AI either removed important language or edited it in ways that didn't make sense to you so it's never going to be as simple as plug and play, it's always going to be sort of this partnership and having that humanity come through at least in my very humble opinion. Oh and then, oh this keeps skipping a little bit, sorry.

[33:11](https://youtu.be/2Lk0MTXzbck?t=1991) There we go and then just a couple of short thoughts on like what it means to be a government employee or work with governments. I think Zach or Zane said this in the chat but like no one's an expert on this yet and neither are organizations so policies like even in Colorado, policies around AI are moving so quickly. Please make sure that before you like sign up for an AI tool using your government ID that you do some investigation with your technology organization to see what tooling has been approved and vetted and just to see you know like what are the guardrails that you're working in and then maybe if the guardrails are up you can certainly do that exploration in your personal life and start to bring some of these policies to your organization could be another way to play it. And then you know I think we've talked about this a little bit today kind of as you know I think we've talked about this a little bit today kind of as a thread throughout but if you are using AI tools you really should be participating in conversations around AI ethics and that you know goes from the artists who are really having rallying comments around their art being used in ways that they haven't consented to or could be thinking about things like AI use statements and you can feel free to rip this one. This is an AI use statement I used, I co-wrote with an AI so we're getting very meta here but those are a couple of prompts that maybe you can start to explore around concepts around AI ethics. And then finally as someone who works at the intersections of government and technology every single day of my life I just really like these questions and I sort of wanted to to end on these but is as with any technology if you're evaluating tools start to kind of get in the habit of asking these questions. Is it free? Why or why not? Is it open source? Is it accessible? Is it excessively built? Who is it accessible to? Both from like a digital accessibility standpoint but

[35:14](https://youtu.be/2Lk0MTXzbck?t=2114) also you know financial accessibility. What does that price point mean? Who can and cannot use that tool? And then these last two that we talked about quite a bit. How are they using my data? How am I becoming part of this AI tool? And do I want to be? And then finally like what is that source material? So we're only as good as sort of the data sets that we're building off of and we're learning from that tool, the AI tools you're learning from. So it's good to be skeptical about that source material. And that's what I got. So I would love to talk with folks more. Yeah Hannah, I'll hand it over to you to kind of walk us through what happens next.

**Unknown speaker**

[35:58](https://youtu.be/2Lk0MTXzbck?t=2158) Great. Thank you so much Abigail. Let me just turn this off of recording.

* * *

Original page: https://innovate-us.org/using-artificial-intelligence-to-de-jargon-government-language · Licence: CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/), as stated on innovate-us.org