 Hi everybody, thank you for joining this webinar. My name is Aretha Simons. I'm the webinar producer here at TechSoup. Today, we're going to be talking about artificial intelligence AI, aka, there's a lot of hype, and this isn't just about hype because it's already reshaping the way nonprofits carry out their mission. Today, we're going to talk about AI for nonprofits and what you need to know. On behalf of TechSoup, we want to thank Microsoft for the grant funding to help make this webinar possible. So thank you Microsoft. Before I introduce our speakers, I want to let you know this is being recorded, and you'll get the video replay within three business days. Currently, everyone is on mute, but if you have a question, I'll show you how you can engage today. Feel free to use the Q&A feature, but my TechSoup colleagues, they're here to help in the chat section. So if you have a question in the chat, we'll make sure we can try to get to all your questions as quickly as possible. If you would like to use the little caption, just go ahead and click on those three dots at the bottom of your screen, and then when you click on those three dots, you'll see where it says turn on live captions. Now, I'm going to turn this over to our two speakers from Roundtable Technology, who have a very unique slide on how they're going to introduce themselves. So over to you, Joshua and Destiny, and thanks for being here. Thank you. Okay, so hello everyone. I am Destiny Bowers of Roundtable Technology. I am our Chief Information Security Officer, and also a cloud strategist, and I'm really glad to be here today to talk a little bit more about AI and frogs. I am Joshua Peske, three CPO at Roundtable Technology, which I'm happy to explain if anyone wants to ask, but otherwise I'll just move along and say I work closely with Destiny and many other wonderful folks at Roundtable, and that we are so grateful to TechSoup and Microsoft for having us here to talk about AI. Okay. And so on today's agenda, we're going to talk a little bit about what Lillipads can teach us about AI and exponential growth. Nonprofits in AI and why now? Why is this such a big thing now for nonprofits and ways that is being used? And of course, an overview of how you can get a little bit more learning surrounding AI. And what we're going to be doing is we are going to be taking our generative AI order. So in the chat, you're going to see a link to a Microsoft form. Just fill that out with something you would like for us to generate. And by the time we are done with this presentation, you are going to have either a generated image or generated text. And we wanna show you in real-time live how this all works. So go ahead, drop your thoughts into that form. We'll pick a couple of random ones and we will go ahead and generate things for you. Dusty, just wanna be sure we're not gonna generate all of them. We're just gonna pick some to do some live demos. Just one or two. Time permitting. Okay. And so before we get started, we wanted to throw up a little poll. And the question is, would you get on this ship? So imagine you have a chance to hop on a magical ship that's set to sail towards the amazing utopia where earth is wonderful, it's healing, healthcare's in and for everything's just top notch. And you got a 10% shot at absolute smooth sailing, an 80% shot of you might get a little seasick, you know, it might be a little bit of a rough ride. And well, frankly, 10% of you're just gonna go right into the iceberg and that's that. So the question is, are you going to get on that ship with us? I'm guessing there's a poll up somewhere. There is, it's in the chat, Destiny. So that's where it takes up our teams. And the responses are spread across but the dominant response so far is yes, but with some concerns, but we've definitely got stuff every. What is your answer, by the way, Destiny? I'm already on that ship. I mean, at this point I'm kind of at the helm. I may not be the captain, you're the captain now, but. Yeah, we have someone asking what were the percentage of the consequences again? I think it was 10% chance that we're gonna sink into the ocean and 20% chance we make it there smooth sailing and like 70% a bit of storminess, but we do get there eventually. Yeah, and I would say that there's, yeah, people are, and I think you all know what the ship is. I think we've kind of, you know, led a little bit on that one. So let's go to our next slide and I'm gonna hand it over to you a little bit about lily pads and whether or not these lily pads are gonna help float the ship or sink the ship. Yeah, so why all the frog images? By the way, pretty much every image you're gonna see other than Destiny's headshots has been generated by AI. In most cases, Dolly 3, which is a new product that actually is available through Bing Chat. And if you're wondering why we have all these frogs, well, it's because we're gonna talk a little bit about lily pads. And I'd like all of us to imagine for a moment we're all living around a beautiful lake and it is May 31st and our lake is entirely pristine. There's not a single lily pad weed anything to be found, okay? And on June 1st, all of a sudden one lily pad appears. And it's a funny little lily pad in that it's going to double every day for the month of June. So it's like the exponential June lily pad. So on June 2nd, we've got two lily pads, June 3rd, we've got four lily pads, June 4th, we've got eight lily pads and so on until, and this may surprise some of you who are not familiar with how exponential growth works, we're over half a billion lily pads on day 30. Now, first question is kind of like a little middle school math riddle and don't feel bad if you don't get it, but if the lake is entirely full of lily pads on day 30, on what day is the lake half full? And many of you will quickly get that, of course it is the 29th day because if they are doubling every day, then the day before it is entirely full, it must be exactly half full. A trickier question is what is the first day that the lake is more than 1% full of lily pads? What is the first day that this lake is more than 1% full of lily pads? And you can do the math. I encourage you just to go with your intuition on this and people can put some numbers in the chat if you have any guesses. What day is the first day that the lake is more than 1% covered? The answer is actually day 24 by which time we're over 8 million lily pads, but still only 1% of the lake is covered in lily pads. So most of us who are living on this lake at day 24 may not have even seen a single lily pad yet, even though they've been growing and doubling for 24 days, and in a mere six more days, the lake is going to be entirely full of lily pads. What does this have to do with AI? Exponential growth does not feel intuitive to human beings because we move through time in a linear way, right? We have May 1st, May 2nd, May 3rd, May 4th. We don't have May 1st, May 2nd, May 4th, May 8th, May 16th, May 32nd, and May 64th, right? So we're used to linear growth, which is the blue line that you see there. But the funny thing about exponential growth is that it feels like linear growth for quite some time until the numbers hit a certain size, at which point it shoots straight up and things start to happen very, very fast. The lake goes from not that many lily pads, less than 1% covered to entirely covered in lily pads, all right, in a mere six days. What we are saying to you is we think that we are living through that inflection point right now, and this is why over the last year, people have heard so much about AI and the changes that are happening. You can see from these graphs that in just the last five or so years, after over 20 years of development of these systems with handwriting recognition, speech recognition, image recognition, reading comprehension, language understanding, image development, okay? These things only in the last five years, these AI systems are suddenly outperforming humans, which is that bar, the benchmarks set to zero at the top have just started outperforming humans in the last five years. Of course, they're not going to stop getting better. They are going to continue to get better and they're going to continue to get better at a faster pace than they have been, all right? Let's look at what this actually looks like, all right? When we take an image of a frog, if you look in the upper left, you'll see March 22nd, right? A year and a half ago, AI systems could barely produce something that looked like a frog, April 22nd, April 2022, still not much looking like a frog. Now November, right? About a little less than a year ago, started to look a bit like a frog. June 2023, look at that next to the actual frog. Next to it, one of the few non-AI generated photos you will see in the presentation today, all right? And then look at a frog with Dolly made just a week ago, all right? Doesn't necessarily look like a real frog, but boy is that an amazing image that was generated. So I've talked to you about why I think AI is so important in general right now and why it's kind of feeling like it's all happening very fast. Destiny's gonna talk to you now about why we think it matters so much to nonprofits. Yep, thank you, Joshua. And this is really important. When anything new comes out that has such a great impact on the way people work and the influences that can have in general, that there's a group of responsible minded people who participate within the process. So on the next slide, what we are hearing a lot about in the news, I think is a little bit of the fear side of AI, which is the, well, it's gonna take over my job. And I think their phrase that's been thrown out there is that no AI isn't gonna take over a job, but the people using AI might. And then also concerns over, oh, Skynet is just gonna take over the world and get rid of human beings. So there's a lot of fear and concern out there. And how are we going to keep the people involved in the artificial? And that's one of the things that we'd like to point out for folks on the next slide is that you are already using AI in a lot of ways. And when you look at the things that you've been using AI for outside of like the recent versioning of the generative AI, you realize that like, hey, this isn't taking over, this is actually a helpful tool for me. And that's what we're also here to help placate people on is that we want to show how this is going to be something to assist you. But the way that you need to participate with this on the next slide is bringing that responsibility, that ethics and everything into the industry. As nonprofits are great organizations who are able to focus on mission, who do great advocacy work, who can bring together coalitions that having that influence, the way that AI is being designed, the information that goes into how it's being used, what kind of changes and really being an influencer throughout the AI industries is really going to matter. And I think that it's something that nonprofits really excel at in the way that they can do things. Because right now, things don't necessarily feel like they're in your control, that there are things that are going on with AI, there are things that going on with the way they're using, the way they're taking data and information that is outside of your sphere of influence. So I wanna talk a little bit about circles of control on the next slide. And this is where in our first ring of concern where these are things I'm worried about. These are things that we don't necessarily have control over and this is very worrying to people. But what you also will have on the next ring is your influence. This is where you can as individuals and as nonprofits participate in the process through advocacy, through collaboration, through joining together in coalition of all of the different industries out there and being supportive of this and also putting your voice into the way AI is being used and developed. And the things that are most certainly under your control would be how they are used within your own organizations, how you help develop the skills, how you put guide rails and other information surrounding what tools should or shouldn't be used and how they're used and also the education and things like that within your own organization. And that's why we really do recommend that on the next slide as a strategy for nonprofits that you start thinking about having policies and guidelines surrounding the use of AI within your organization so that you're vetting the tools you're looking at them and you're knowing how information and data is coming in and out of your organization and that you're not having any risks for privacy or other intellectual property concerns or any ethical concerns and how you inside of your organization are using AI and AI related tools. So speaking of AI related tools we're gonna throw up our next poll which is what have you made? And I know we threw some of this out in the chat already but what have you made with AI? Are you doing texts for emails, poems? Are you translating things like in videos, images, data visualization, translation of data, slide decks, audio, video? So do a little vote in the poll and anything else that you might be using it for let us know in the chat because we really do want other folks here to see how this is being utilized. We did have a percentage of people who said they weren't using it at all maybe that's cause you're not sure how to apply AI where can I use it? And so that's one of the things that I think you can learn from each other again the coalition of learning is what are good things that we can use AI for within our nonprofit? A lot of SEO. Just under half of the folks have used AI to generate some texts so about 45% so far about 17% have used it to translate something into another language. Under 10% about 7% have used it for some kind of data visualization. 14% for image making, 7% to make a slide deck. So there's definitely some folks that have used this but well under half for everything except text generation. Newsletters, elevators, speeches, appeals to lead poems, I love it. Yeah, there's some really great things in the chat, grant writing. Yeah, and I think what this does is it allows people who may not necessarily have that skill set 100% like it becomes an aid. It becomes something that will help you into an arena. And I think you have a thing that you used Josh with like not everybody is a data analyst but with AI you can be. Yes, well we'll be showing a video about that. So yeah, it has both as so many things with AI, right? There's tremendous opportunities here, also some pretty serious risks but a lot of great opportunities don't come without risks and what we're here to try to help everybody do is realize some of the opportunities and really manage and mitigate those risks. Yeah, so I think there might be a little bit about test driving this. Yeah, well let's see what responses I've got in so far. So I've got some. All right, so we're gonna do a little text first. So we've got- Real live demo, let's see how well it goes. Yeah, so live demo, always a challenge, right? So let's head over to Bing and let's share this tab instead and everybody should see now my Bing screen and I can make it a little bit bigger. So folks- And this is available now for free to everyone just through your Edge browser. It'll even give you a little pop-up when you first log in about utilizing your same tab. You can use Chrome or Edge so you can either use the Chrome browser or the Microsoft Edge browser to the best of my knowledge does not work in other browsers and you need a free Microsoft account that you log into to use this. And one of the things we'll talk about is actually the enterprise version of this which basically is the same functionality right now but has additional privacy protections is available for folks who are using Microsoft 365. So I'm just pretty much gonna paste these in verbatim. So someone is here said please generate a safety briefing for being on a dusky boat and also scuba diving and shark tagging via that boat. Join us for our Christmas concert during Sunday worship. Interesting, okay. Those seem like two different ones. So I'm gonna do, I think I managed a tag. So we're gonna safety briefing for being on a dusky boat. So let's go ahead and do that and let's watch our AI generate some text on that. And while it does that, so it's one of the things about Bing that's a little different from the free version of chat GPT that I think many people use is that it has access to the internet which the free version of chat GPT does not. It is only trained on data going up to the end of 2021 doesn't know anything about that. And often will although less of these days will make up sources for things as a lawyer famously learned a few months ago. What you will notice that Bing is doing is actually sourcing websites. And so you can actually go and quickly validate those sites and just make sure that they actually exist and they're actually good data. But you can see that this does really a pretty excellent job whoever put that in. So let's go grab another one here. So I'm just trying to get my responses. I'll see if there's a better way I can do this bear with me just a moment. All right, so I've got some of those. So let's an email politely ask, I mean, everyone can use this. Okay, an email politely asking people to clean up the microwave after they use it AOL, LOL, sorry, please grab friendly and humorous. Yeah, let's see if I can spell humorous live. Did I get that? I don't know. All right, and let's see what it does. So I added the only thing I changed of the prompt the person gave me is I just added friendly and humorous to try to give it a little bit more guidance. Now, if you come to the class, which we'll tell you about at the end of this, we're gonna get into what's referred to as prompt architecture and all the different prompt strategies that you can use to try to generate better outputs from AI tools such as Bing Chat, which we're seeing here. So if you wanna kind of learn all of the tips and tricks and best strategies and things around prompting, that's going to be, we're gonna have a whole class dedicated to that in the course that we'll be sharing with you at the end of the pre-session today, okay? So we'd certainly love to tell you about that. What do you think, Destine? Should we do one more? Yeah, I think we have time for one more, yeah. Okay, so let's see. I got another text. What's a good one? I'd like to create a, let's see. All right, well, this is, I think everyone then, okay, a year-end campaign theme that will generate 50,000 in revenue for my nonprofit. I think we have to pick a nonprofit. So I'll just make one up. So- How about the QT Collective? Help me, yeah, exactly, right? A year-end campaign theme that will generate 50,000 in revenue. That will generate 50,000 in revenue. And one of the, well, Jocelyn's doing this, one of the great things, as you mentioned in the prompt writing is that we like to think of these a little bit as the unpaid intern where you can go back to it again and again and ask it to reiterate or make changes or things like that without getting annoyed. Like, are you asking me to do this again? And you can just keep rephrasing the question saying whatever adjustments that you wanted to make until finally you're getting the output that you want. So you have your first draft and then you can keep iterating and iterating without worrying about someone going, I can't believe this is the 20th time you're asking me to rewrite this. By the way, I think our kitty crochet collective destiny must have been out there enough that Bing knows about it because it seems to be familiar with the kitty crochet collective mission. I wonder if we actually made a website for it at some point and it actually did find it. We've got our first shifts, which of course, yeah. So as you can see, it does give, and you can have it, right, Jocelyn? So it is giving us some great output here of ideas. And when you see other learn more links, you also have ways to iterate there if you want more information, things like that. So you can expand upon the theme as it were. And so now I think, yeah, so let's, I think two slides forward. Yeah, right, so back to you, Destiny. Yeah, so, and that's what I was saying while we were doing the demo is that this really is your writing buddy. A lot of people suffer from blank page syndrome where you might have an idea, but you don't know how to get started or maybe you don't even have the idea and you've got to get started. But this is the great thing around a lot of these tools is you can throw some initial very basic concepts at it and have it do that first iteration. And then you can go back to it again and again to make changes, make adjustments, whatever you need. And then it can help be your copyright editor. So it's your writing assistant. It's your copyright editor. And it can also have SEO. So if you're trying to put something out there that you wanna make sure has good metrics on it, it can do all of that as well. The one thing that we always say though is that if you are not an expert within the topic that you're having, they're having it right on is to make sure that you fact check anything that any kind of references or any other statistics or things that are in there are indeed factual. So you do need to know sometimes a little bit about your topic to make sure you're doing your due diligence. If this is just creative writing without anything like that, then you don't need to do that. But also with certain ones of the generative chat ones is you can give it personas. You can say right in the style of or this is this kind of audience and you can even iterate how you want it. So you could say, this is for a nonprofit audience of this type of work or what have you. And you can adjust the tone in the way that these chat bots write for you as well. So there's a lot of things you can do not just in the way you write the prompt, but also in the characterization that you give it as it creates that content for you. And I think we are gonna take a quick look now at the other thing that I find incredibly helpful myself. And I think we are all in, I was just gonna talk for one second about the note takers on the previous slide. And that's just, this is a tool that I think a lot of us might have seen or might even use. And now you have built in or available as an add-on to a lot of your video programs, the ability to have note takers added in. And these are really helpful so you can be mindful and pay attention to what's going on in your meetings while somebody's out there taking notes. So at the end of it, you say, well, what output do we, what outcome do we have here? What was I supposed to do? You will have someone who has taken those notes for you, which can also be helpful in translation to give people text versions of whatever happened in the meeting, which can also be helpful for record keeping and things like that. Just as a little caution, be wary of that they have access to what you're talking about. So make sure the conversation that you're having doesn't have anything sensitive or private in it when you're including your note takers. But now we get to talk about pictures. So if you loved our frog before, you're gonna love what we're gonna do here because a lot of us are frustrated artists. I know personally I can't do anything beyond a stick figure. And there are so many things that you do on the day to day that really benefit by having images in them, thousand words, pictures, we've all heard it, but maybe you are not so good with the drawing. And this also can create the content but it's helpful to generate even life like people. You can generate fake people. So if you need to have people for your website or a press release where you don't have the ability to get that kind of input or you don't need or don't wanna have people have to sign forms that allow for, particularly if there's children or things like that involved, you can AI generate people now. And I think they're starting to get hands a little bit better than they used to. So there's a lot of ways that you can use your generated content on the image side. And we're gonna use it to play a little game. And so our little quick game here, I think one more slide is throw this into the chat. What famous song title verbatim is this image displaying for you? I think this one I feel like is a little bit of a giveaway. Purple rain, they got it. All right, so next one. Okay, and the next one, what's this one? I think this is a, at this point, I didn't even know how many decades. Stairway to heaven, they got that. Okay, last one's a little bit harder, but we'll see how many people can get the exact name of the song. And I will say though, that I did not know the exact name of this song either. Everyone who's getting it, yeah, it is, everyone's putting in lime in the coconut, but it is just called coconut, which I did not know. I always assumed it was lime in the coconut too. So we've learned something in music theory today, generated by AI, so now. Yeah, so I'm trying my best to make images as fast as I can. This is the, the image making is a little slower. So without, you know, we don't want to keep people, you know, here too long. So I've got a couple that I preloaded and then I'll try to get some others just to show people how we do it. We'll go back off that one. So we're going to share this. So let's see. So here is a, someone requested a purple dragon driving a little red Corvette with his tongue hanging out. And so here are the images that we got from that. Not too bad. I don't know why we got this little thing on the end of the tongue there, but and here we got a co-writer with our dragon. It's very cool. We didn't ask for that, but we got it anyway. Another image that we got asked for was someone, a person playing a drum, Congo style. So I've got those in there so people can see that it, the way Bing and Dolly three works. And for those of you who haven't used Dolly three yet, which is the new Dolly version that is baked into Microsoft Bing, it is exceptionally good. This is a world of improvement over Dolly two, which was the previous version. So let's look at how this works. So what I'm going to do is grab another one. So I got, I'm trying to pick a fun one here. Oh, this is a good one. Okay. So we have someone who said, I would like an image of, I'll type this in. So please make me an, by the way, I'm one of the people that likes to talk nicely to the AI, you know, just so that they think of me as a nice person. Very polite. So when the AI overlords can speak over, they will be nice to Joshua. All right. So what you do is you plop it in and it will kind of chug away for a while. It's, this is a little bit of a slower process than the text generation. So you'll see this, your image is generating and it'll kind of chug away for maybe 10 or 15 seconds and hopefully soon we'll get an image of a tuxedo, we'll actually get three or four images of a tuxedo cat astronaut on a national geographic cover. And one of the things that Dolly three does is that it actually spells a bit better and you'll see that you get these weird artifacts, right? So here we've got this random kind of Instagram female face in the hood of the hell that the cat is holding. It's kind of random. You'll notice that the spelling of national geographic is pretty good, much better than Dolly two did, but there's still going to be misspelling. So you're still going to get two E's in geographic. This one looks pretty good, right? That's pretty solid, but you'll still see some, you know, gobbledygook words over the national geographic and it did still misspell geographic, but better. So anyway, that hopefully gives folks an idea. Destiny, do you think we have time for one more image? Should we try to do? Well, actually what I wanted to do for a moment, I did want to address something that has been repeatedly asked in the chat about the copyright of AI generated images and the use of AI generated images. I need to answer that. Yeah, do you want to take an answer that, Destiny? You want me to do that? Yeah, I mean, I was just going to say, you know, right now things that are generated by AI that are not based off of other people's work are not copyrightable. It is basically open. Things that are based off of other people's work is being heavily litigated now in the court system. So if you generate something using AI, not based off of, you know, make it look like this person's artwork, you cannot copyright that image at this time. So it would be in the public domain. Therefore, if you choose to credit, you could say generated with AI or generated with Dolly or something, if you want to, so people understand that this is not real. And I think people are a little bit, when we talk about the ethics of these things, people are being a little sensitive over what is a real image versus what is AI generated. So you might want to indicate that, but no, you cannot copyright at this time, AI generated images. Yeah. Josh, would you want to add that? Yeah, I would just say one way to avoid potential copyright claims. Using Dolly 3 as opposed to mid-journey, they, Microsoft and OpenAI are putting more copyright protections and controls into those than mid-journey is and certainly more than stable diffusion, which is another popular tool. People are asking if these are free, if they pay other tools. We'll get more into specific tools and strengths and weaknesses of different tools in the course. We're not going to do too much of that today, but I will say what I am showing you right now is free in Bing. I'm doing all these images in Bing live right now for free without having paid any money. Another one we'll just show you real quick. Someone had asked for a flourishing tree that casts a shadow where the shadow represents the tree it was withered and uncared for, which is pretty tough. Didn't do great at that one, I don't think. So that's an example of where you need to sort of explore with different prompts and see if you could get a sort of better prompt out of that. But one way to avoid copyright would be to, let's say not ask for something national geographic in the image and not ask for something that looks like a Banksy painting, like a living artist that still has copyright protections clearly, you can try to avoid it that way and then also avoiding using tools like mid-journey or things like that that might have it. So there's a lot there and someone just posted a link in the chat to around some of the copyright stuff. So yeah, certainly with everything. We're having in the chat about ethics, that's a big part and that's why we're saying it is an important time for nonprofits to get involved in this because of the ethical concern. It's not about the money and the generative income from AI, it's about the ethics behind it and that's why these are great voices to be brought into the conversation about AI. And one of the sessions of the course is entirely designed to talk about ethics, safety and all these kinds of views. And if it's not clear, Destiny are trying to say it every time we talk about it, there are great opportunities here at Great Risks. This is why we want nonprofits involved because nonprofits are thinking about these things in a very active way and we want your voice in this space but you can't have a voice in this space if you're not using the tools and understanding what they're capable of. All right, so we're gonna show you a kind of cool capability that Bing image can do which is it can do this in the other direction. Okay, so Destiny and I have a webinar, free promo here coming up next week called Scary Stories with Roundtable and here's an image that I just grabbed for this webinar. So I'm gonna show you a quick video here and there's no audio with this so I'll just talk it through. And what I did is just uploaded this video to Bing chat. All right, there's a little image button you'll see there and you can just upload it and say, tell me about this image and look at how fast Bing is able to go in and tell me everything about them. That is just from uploading the image it was able to tell us everything about the webinar, everything about tabletop exercises, even the tone of what we're trying to go after with the webinar, it is astounding the kind of information that you can get just by uploading an image and asking some questions about it. And there are protections in there in terms of faces and things like that that are certainly there. Now this we're gonna show you this is not a Bing capability. We're gonna show you two different tools here. One is Claude by Anthropic and another is ChatGPT, the paid version. Okay, and this is way to analyze data. So my colleague, Dustin's colleague Kim Snyder is gonna talk you through this. The volume's a little soft. So you might have to turn your volume up a little bit and this video is about two and a half minutes long. With Claude, I'm able to upload a file. In this case, I'm uploading a Word document file and it's a New York state privacy law. So a little complex to read. And I can upload it and then ask it questions that I have about that law. So once I upload it, ask it its questions, takes a little time to process really quick and then it provides the answers. I can then also have some follow-up questions for it and keep asking it questions about this law until I understand what I need from it. And in order to be able to double check or verify, I can also, I do ask it to site page numbers, as you can see. With Claude, I'm able to upload a file. In this case, I'm uploading a Word document file and it's a New York state privacy law. So a little complex to read and I'll cool. So it allows me to upload data files. This is a CSV, this happens to be public data employee records, but what it will do is it will read that data, confirm a preliminary analysis of it, ask me to confirm it, and then I can provide it with a plan or it will suggest a plan and then I can ask it to modify that plan and do what I need it to do. It will also then come up with some high level findings initially from the data, ask if I wanna explore further. If I say yes, then it can also do visualizations. And in this case, it's generating a lot of different kinds of visualizations, allowing me to look at different factors in this data. For this reason, it's great for exploratory data analysis. It's also capable of doing its own data cleaning, which I find amazing. And it can stop, when it stops to see, even this case, it's a field that's not been, that's formatted as a string. It recognizes that, it then converts it so that then it can generate this next chart. Okay, it can also look at multi-factor analysis. So in this case, it's the box and whisker chart looking at the effects of two different variables. And the advanced data analysis tool can also then generate an executive summary for me, or it can generate a list of prioritized findings that I could take to this hypothetical HR group. With Claude. Go ahead, Destiny. Ah, well, I was gonna say, I'm not sure if we ended up getting the poll for this one, but before we talk about AI video and translation, we do want to, in our poll or rather in chat, throw out a question. What are we naming our froggy director? Should he be Federico Fraguini, Pedro Totemovatar, Werther Hertzfrog? Or do you have your own idea for our frog director? No, we do have the poll. We do have the poll. All right, yeah. Federico Fraguini, Pedro Totemovatar, Werther Hertzfrog, or you could write your own. Hermit is one, got a vote for Werther Hertzfrog. So far, Federico Fraguini is in the lead here, so. Oh no, I think Werther Hertzfrog is bringing up a close second though. Yeah. So we're gonna- Froggy script is not for this frog yet. I'm gonna play a video. And by the way, I made a mistake. I made this earlier this week and I made an error here. So I'll offer a little prize of giving you a shout out if you're the first person who identifies what error I made. It's a fairly significant one. But this video demonstrates some language translation capabilities that exist in AI systems right now. And so I'll just play this really quick. In this video, you are going to see me speak a few different foreign languages, provenienti da diverse parti del mondo. Sembra che parlo fluentemente queste lingue. Franchement, je ne parle presque pas anglais. Pero in este video, me veras hablar inglés, luego español, francés, italiano, y finalmente, hindi. Apen bachau me, dhara pravaha yaswadeshi bolne baale ho. Me, dhara pravaha se bol raha ho. So, what'd you think? Yeah, people got it. People figured it out. Yeah, yeah, someone pointed that out. It just shows how bad I am at languages, right? Like, I don't even know what I was saying, speaking. And what's interesting about this, for those of you who are in the New York area, there's a little bit of controversy right now over our mayor using this exact tool to do calls and videos in other languages and whether or not the portrayal that he speaks this language is something that falls under a question of ethics versus just wanting to try and get the message out in as many languages as possible for folks. So sometimes you never know what might turn up a little controversy until it's out there. But if you are giving talks or presentations where you do have multi-language speakers, this is actually something I believe is very helpful so you can really communicate with everyone without language being a barrier. All right, so Destiny, what are the next steps here? Well, that is a great thing. So the first thing, as I said, to be part of the conversation, you kind of have to know how to speak the language, right? Speaking of language. So get out there, use these tools, test these tools, try them out. So the more you're informed about them, the more you can participate in how they're going to be developed and everything else throughout the world. And then also internally, have a strategy for your organization. You do want to give some of those guide rails to keep everybody on the right track when it comes to the utilization of AI so you don't run any risk of copyright infringement or in putting out sensitive data into another learning model when that was not intended and then learn more about it. As we said, we are gonna have the course and I think Joshua has a very interesting AI generated nightmare to tell us a little bit about that. It's not anything like me, Destiny. Hello, this is Creepy AI Josh Frog and I'm here to invite you to a four-part course on exploring AI with non-profit tools starting on Tuesday, October 31st. So the first session on the 31st will be introduction to AI and generative AI on November 7th, one week later, we'll do ethics and principles of using AI responsibly. On the 14th of November, we will be deep diving on prompt engineering and we're gonna wrap things up on 1128 after taking a week off for Thanksgiving week here in the US and we're gonna have BingFest, AI use cases for non-profits. We would love to see you there so I hope you sign up for the course. Oh, and I suppose I should say one last thing, ribbit. Do I owe everybody apology and apology for that, Destiny? I feel like you do. You do, absolutely, absolutely do. So for those of you who tuned that out because you were a little scared of it, here's information on the course which does start on Halloween, which is why we want to make it a little creepy and then we'll be continuing one time a week with the exception right around Thanksgiving, which it will then pick back up for introduction, ethics, prompt engineering, and then kind of a BingFest of how Bing can help you with all of your AI projects. All right, and I think that we're ready for Q... Go ahead, Rita, why don't you take it from there? Yeah, that was awesome. I love that ribbon. There was a lot of lacking emojis in the chat, so awesome. I'm gonna, before I wrap up, I have a few questions I think, probably deserve some answers here. One very important question is, how do you properly credit work from AI? You want to take it, Destiny? You want me to do that one? You could take it. Yeah, I think that there is, this is the thing why we're encouraging organizations, Rita, to really come up with policies for their organizations. Here's the short thing I will say. Right now, 45% of the people in this session said they're using AI to generate things, and that's probably low, quite frankly. And if you think about, I'm guessing, if anybody here works for an organization that has provided you with an AI use policy, guidelines, training, has communicated to you in any way, this is what we do want you to do with AI, and this is what we do not want you to do with AI, then you will be one of the first organizations, I have heard of doing that, that isn't an organization that Destiny, myself, or colleague Kim has been working with. And by not having any kind of AI strategy organization, in our view, you are accepting most of the risks and missing out on a lot of the potential benefits. And one of those is, hey, if you use AI to produce something that you're going to share with someone outside of our organization, what should you tell them about what you did? Should you tell them this was like 60% me, 40% AI? Should you tell them I used AI to help this, or should you not say anything? There aren't laws about this right now for the most part. So that's why you as an organization have to decide what your policy is, because I think we all see how fast regulation gets through our Congress. I mean, I'm sure Congress is going to pass some sort of AI regulation. Oh, no, wait, they're not doing anything right now. They're just voting a whole bunch on something that whether they could vote on other stuff. So yeah, so I don't think regulation is coming down anytime soon. So it's kind of on us to figure out what we think is appropriate. That's my answer. Thank you. Next question is, are there any free AI website for images? The one I just showed you. Yeah, so I can, we can go over to Bing and I can share this tab. We can go to chat and making images here. So I've got another one that someone, while we're doing the next, you know what, I'll just keep making images of stuff while we go through the Q&A unless anyone objects. So I'll just keep sharing that screen. And here's another one that someone asked for, but this is free, what I'm showing you right now. Sign up with a Microsoft account, 100% free Bing Chat with Dolly 3. Mid journey is 10 bucks a month. And I think that there's, I think they just added Dolly into the paid version of ChatGPT as well. And right now Adobe Firefly I believe is also free. Okay, awesome. Another question, which note taker AI tool do you guys recommend? I don't think that we would officially recommend any of them. I can tell you that we use Fathom. We have reviewed fathom.videos security and privacy policies and are comfortable with them. Although we do delete our Fathom recordings after 30 days, Fathom doesn't offer a means to auto delete them, which we would prefer. Otter, O-T-T-E-R is a good one. And I've actually used their digital trust page as an example of, I wrote an article on LinkedIn about digital trust and transparency and Otter's was the example I used. So they give you, they're very transparent about everything even to the point of if we get a subpoena for your recordings, here's what we will tell law enforcement and here's what we won't, here's what we have, here's what we don't. So they're really, really good. Another one, we use it as fireflies.ai. No, I don't know much about it. Go ahead, Dustin. I was just gonna say though, is that like any other tool, anything that is creating recordings, transcripts of things in video chat, you do want to be aware of how, what kind of data you're having in those conversations, should it be recorded, the legality of third party states and recording and everything else. So we normally will also make sure that anyone else in the meeting is comfortable with this information being captured. So that's another thing to consider, again, going back to your having a policy and how you do things internally is to say, when is it appropriate to utilize these tools, what is the retention period for the information that you're recording in the transcripts and how will those be used. So it's one of those things in the background of every tool that you utilize. And I see a lot of these questions in the chat of like how do we make sure of or what should we do and you are, there is, I'm seeing a big desire to build out those guidelines. And that is that discussion that you wanna have within your organization to help establish these things. Awesome, thank you, Destiny. Another question is, do you need to anonymize any data you enter into AI tool? Yes. Yes. Well, well, well, yes, yes. There were gonna be the exception, which Joshua will talk about, but if you are using free tools, absolutely. Joshua, you can talk about the paid tools. So if you're a Microsoft 365 organization, meaning that's your primary email organization, and you're on, and this gets technical, but if you're on an E3, Microsoft E3, Microsoft E5, or Microsoft Business Premium License, if you don't know, you can go ask. But here's a quick way you can find out. If you just make sure you're logged in with that Microsoft account and you go to Bing Chat, if you in fact are on one of those licenses, you have Bing Chat Enterprise, you have it already. And you can tell, because it'll say it's Bing Chat Enterprise, we have a big enterprise there, you can't miss it. And that is protected by essentially the same data and privacy right now in controls that all the other data in your Microsoft tenant, your Word documents, your PowerPoint slides, all of that stuff is protected by now. That does not equate to you have cart launch to take any data and dump it in there and start doing analysis. That is something that you should consult your AI policy at your organization, which you don't have because your organization hasn't made one. So that should happen. But that is a promising development and I expect to see a lot more of that. Yeah, and we're gonna see that coming out and other ones I know that Google's developing it. So basically any of your big players in the cloud space, it's already in within Salesforce. So Salesforce has, I believe, Einstein, which is also covered by your agreement with Salesforce. So these are out there, but just because it's available and it does have data protection, you still wanna review your terms and conditions, your privacy policies of that particular product to still get an understanding of how that's being used internally by them, even if it's not being put out into the larger language learning model that the public can have access to, you still wanna have a solid understanding of how it's being used within the vendor that is controlling that product. So there is gonna be some due diligence if you wanna put data in that is not being anonymized, do your due diligence on how that's gonna be handled before you put it in there, otherwise anonymize, anonymize, anonymize. Okay, awesome. Here's a question. The video shows data analysis, but what about the protections over that data we use or pay for programs to keep the information safe and less accessible? So what about the data protections over that data? You wanna handle that one, Dustin, you want me to take it? Well, unless I'm misinterpreting the question, I feel like we did cover that a little bit previously, which is that if you're putting in data into a free tool, then there is no protection that is going in there. And if you are not anonymizing that data, then you risk exposing that data and it is not being protected in any way, at least if that's how, if I'm understanding that question correctly. Okay, hopefully they can watch the replay and gain some more insights from earlier. Here's a great one. Can I use AI to generate a children's book like an illustration? Why not? Yes, and many thousands of people have done that to the degree, Aretha, where people are making children's new children's books for their kids every night on demand. Kid, what do you want a story about tonight? And it writes the story, makes the images and you read that book. Wow. So yeah, 100%. Thank you. But you can't come up here and write it. Oh, good to know. Not yet. Good to know. Someone came out nice, the grandparents in this. Yes. But that's a good thing. Speaking of advocacy, like the ability to people to influence the laws and protections that are put in place, do you want to be able to copyright this information? Do you want it to be public domain? I mean, that's a decision everyone has to make into how they feel about this, but that's also where you can influence the laws that are made surrounding this information and this data and the generation of it is you can take your feeling and you can influence, that's the point of it is influencing our leaders or our voters or our legislators. Rather, I'll pick a word until I get the right one. Someone, Josh, can you see that for me? And then- The people who don't work because they just like to keep voting yes or no on something. But yeah, that's where when we talk about the influence and someone did actually ask, Josh, would you can bring up the slide about circles of concern? I saw that in the chat where someone I think had asked if we could bring that slide back about with the circles of concern where we have the concern, we have the influence when we have the control rather the circles of control slide. There was a request to bring that up. Will we bring that back up, Ritha, any other questions? Yes, yes, yes, yes. And I'm gonna put these two together. Are there any resources or example of internal policies or template or standard for AI policy that we can utilize? What's that? Absolutely. I will tell you if you want it for free, you're gonna have to go find me on LinkedIn and find an article I wrote about six months ago. I'll give you a hint. The title of the article is, should, is it time for your nonprofit to have an AI policy? So you can look up my name, Joshua Peske, you could Google something like nonprofit AI policy and maybe you can find that. Otherwise come to the course because we're gonna share that with you in the course and we have a session we're gonna help you. We want everybody who comes to the course to come away with a policy for their organization. My last name is Peske, P is in Peter E, S is in Sam, K-A-Y. Why are you doing that? Last question, what do you mean by anonymizing? People don't know what that means. I don't know what that means. Yeah, it's not the easiest thing in the world but it's basically getting rid of any data that could be used to identify, generally speaking, a person. So if I'm going to download a list of everybody that attended this webinar and gave feedback because I'd like to have AI chat with it and find out what people thought of the course and interview based on AI. Another thing these systems really get at is talking to qualitative data. So if I have a bunch of words from interviews, from feedback surveys, and I wanna ask questions of that, AI's really good at that if I can load it all in but I can't give all of your names to a system that doesn't have privacy controls. That would be a violation of data sharing laws that is a violation of laws. So that is what anonymizes me. So what I could do is go through and take those fields out, right? Replace them with gobbledygook, replace them with pseudonymization, right? Where there are all other names so I can track identities across but you have to be very careful about that and that can be a lot of work if your data has a lot of identifying information in it. So proceed with caution. So it's really a lot of find, replace, remove names, put in ID numbers or anything like that. So as long as you're pulling out things that are names, addresses, phone numbers, that if I found this information online, where it will say, give me a list of people who work with this organization, you don't want what you've just fed into this to come out. So instead it would be a lot of Fred Flintstones and Barney Rubbles or what have you. Okay, I'm having trouble unmuting myself. Okay, good. Did you want to show that example? Because if not, then we'll go ahead and close it out. We had about a minute left. Which example? The circles of control? Yes. Yeah, yeah, it should be, everybody should be seeing it now, yeah. Okay, awesome, awesome. And just to remind everyone, we are recording this. You'll get the recording within 48 hours. You'll get links, you got hyperlinks on the video slides. You're gonna get the slides as well. If you want to take some of the AI courses, just want to mind you that TechSoup is a nonprofit, 501c3, just like you. And we're gonna offer you these courses at a very, very reduced rate. I think you can take all four courses for $90, which is on heard of. And as I said earlier, AI is not just hype. So I definitely highly recommend you take these courses. You'll probably see me in some of those courses. I want to thank Destiny and Joshua for a great presentation. And all of my TechSoup conference in the background, Matika, Kevin, Vanessa, Shria, Sabah. I hope I didn't forget anybody. Thank you all so much. And again, look for the video recording within 48, within three days, three business days. Have a great day, everybody. Bye-bye. Thank you so much. Thanks everyone.