← School Business Insider
Loading summary
Transcript71 lines
- [00:02]
A
You're listening to ASBO International's School Business Insider. I'm your host, John Brucato. Each week on School Business Insider, I sit down with school business officials and industry experts from around the world to share their stories and explore the topics that matter most to you. Find out what it means to be a school business official and get your insider pass on all things school business. Hello everyone, and welcome back to School Business Insider. In early 2024, I sat down with Brian Cheknicki, Executive Director of ASPO New York, to talk about how he was using AI in data tools to create new resources for school business officials across the state. Since then, his work has only grown Transforming how ASBO NY supports its members, advocates for districts and simplifies some of the most complex financial data in education. In today's follow up episode, we'll revisit that conversation over a year later to see how those early ideas have evolved, what new innovations have emerged, and how AI is changing the way associations like ASBO NY serve their members. Brian, welcome back to the podcast for what I'm guessing is maybe the 15th time now. I don't know. How are you?
- [01:23]
B
Yeah, good. Thanks for having me. Yes, I did stop counting, so I don't know anymore. But I'm very happy to be back.
- [01:30]
A
Happy to have you. And you know, I was prepping for this episode and one of my things was to actually tally how many times you've been on, but I completely forgot. But next time, which I'm sure will be in the not too distant future, I'll have a good number.
- [01:42]
B
Sounds great. Yeah.
- [01:43]
A
And anyway, so as I said, you and I, you were one of the first episodes since School Business Insider really kicked off and I had a great time talking to you about using AI and how much you effective affectionately talk to chat GPT as chat, which I thought was hilarious at the time. But I want to know, so since we last spoke, how has your usage of of AI really evolved? I know you started doing some pretty basic Excel functions and breaking down complex data for, for districts in the state of New York. But tell me, like, what has your, your journey led you? Where are you today with, you know, using AI and what has been the biggest change for you?
- [02:22]
B
Yeah, no thanks. Thanks for asking about these because, you know, we're sort of in the, in the middle of our sort of budget advocacy cycle. Not, not in the middle of the cycle, but sort of in the middle of the prep work for that. And so, you know, this is, this is very, very much on my mind Right now, before I talk about the specific sort of ASBO uses, actually I just, I have a couple different personal uses I've had for it that I think can help help paint how AI can and can't be used in different settings. And I think when you sort of branch out even beyond your work cases, it can give you a sense of how good and not good it can be. And so just kind of in no particular order, it's largely the replaced Googling for me. So I don't search that much anymore for four things because I find, you know, just personally, I find the search engines as they've, you know, become more monetized. There's there, you know, there's a lot more ads and targeted results and less on the. You know what I actually need when.
- [03:29]
A
I'm looking up to scroll half the page to actually get to the first organic result.
- [03:34]
B
Right, exactly, exactly. So just a couple of things where I've used it. I had a. So we have a. We have a camp up in the Thousand Islands, which is in the northern part of New York. And there's. Every spring when we're there, there's bugs that I don't see anywhere else. I haven't noticed them around my house or anything. And so one day I'm like, oh, I wonder if Chat knows what this is. So I took a picture and I said, this insect is located in Cape Vincent, New York, in May or typically in May each year. Do you know what it is? And it returned back an answer. And it turns out it was some sort of river midge. But it's the type of thing that I couldn't. I wouldn't be able to Google that. I don't know what I'm Googling.
- [04:14]
A
Yeah. How would you describe what you're trying to.
- [04:16]
B
Right, exactly. And related to that, I've had some DIY things that I've used it at home. So I've had some plumbing, I've had some car stuff. And you try to Google, you try to look up videos for different sort of instructional videos. And what I've done is I've taken a picture of whatever it is I'm looking at and I've described the situation and it's actually been pretty good in giving me answers back. Oh, yeah, that's a serious problem. Or, you know, here's some suggestions on how to do it. One of the plumbing things I had to. I had to outfit a pipe going into a toilet that didn't have a valve on it. But, you know, I'm not A plumber at all. So I'd never done this before, but you know, I'd say the, you know, that was like 80% chat and 20% the guy at the hardware store helping me out. But just sort of, you know, all these different cases and just two more. And I don't mean to belabor this in, you know, your episode time here, but I think they're good use cases. So my family and I, we went to, we had an ASBO meeting, sort of an executives meeting in California. And my family had decided to come out with me. And last minute we decided to go to Disneyland. Now usually something for, you know, the expense and the time you're going to spend there, you're going to do a lot of time planning. Well, we decided on Thursday that we were going on Saturday. And I'd been there, you know, once 20 years ago, my family had never been there. So I wanted to make the most of it, of it. But not really enough time to come up with it. Not really enough time to, you know, they have, there's contractors that you can work with. Right. Who will plan out your day and all that.
- [05:52]
A
Right.
- [05:53]
B
So I actually used chat GPT and, and I gave it sort of the targeted time. I said, all right, it's the, you know, the third Saturday in August based on what, you know, of, you know, the, you know, how busy the park's going to be, let's plan out the day. And so it, you know, it kind of gave me an itinerary. It suggested when I should use my lightning passes and all that. But then what I did throughout the day is I kept updating it. So I'd say, okay, it's now 11 o'. Clock, we've hit these three rides. Readjust the schedule. And all day long I did that. And I tell you what, we hit every single ride we wanted to do but for the Matterhorn. And that was only because we decided everybody was kind of roller coastered out. So we didn't do that. And every wait was 20 minutes or less, except for the very last one of the day. So I'd sort of used up my lightning passes and I didn't, you know, I couldn't really maximize that anymore. But we ended up in a, you know, being there for 10 hours.
- [06:44]
A
Hours.
- [06:45]
B
We got 12 major rides in plus, you know, plus lunch and all that. So it just, it was a, it was a great, you know, Googling that would never, never get that.
- [06:56]
A
There's no way Google would have helped you with that away.
- [06:58]
B
Right. And, and, and, and Certainly, you know, I feel bad for the contractors who, who you know, are probably losing business to that. They do this as, as their job and they're losing this. But you know, I just, I think, you know, I think that's some of its ability to pull information and really distill it in ways that are specific to what you're prompting it for as opposed to just general questions. I think has been incredibly important in learning how I can then apply it into the work that we do.
- [07:29]
A
Well, if you think back to almost two years ago at this point, 18 months maybe when you and I talked about this, if I am Remembering correctly, ChatGPT was very limited in its searching functionality. So when you asked it something, it was basically going off of knowledge it was trained on, it wasn't actually accessing real time data. We're now fast forward to where we are in November of 25. It's doing real time searches. So I don't know that your experience specific in the Disneyland example would have been the same 18 months ago as it was, you know, a couple months ago, right?
- [08:04]
B
Oh, I, yeah, I think, I think you're absolutely right on that. And you can see it, you know, as it's, when it's doing its work right, you can see it doing the searches. It's, it's getting better now about sending you the source link so you can go back and verify, you know, whatever it's telling you is, you know, is accurate from what it's pulling. But yeah, no, it's, you know, it has definitely become very much integrated into just my day to day routines.
- [08:30]
A
Now I have to ask, are you using the voice function? Are you talking to it or is most of your interaction still typing?
- [08:37]
B
It is mostly typing. I have used it to dictate things and there's sort of two work examples where I've used. I think in the work that we do, we have really prided ourselves on being the originators of the content. We have staff, we have the capacity to write stuff ourselves and so we've tried as much as possible to do that. Where this has been helpful is oftentimes when you're writing something, whether it's a report, an email to members, all sorts of different things is sort of that initial time of sitting down and getting that first draft out. Once all your thoughts are on paper, it's in, it's in good enough shape that you can move around to. And so what I've done, there's two specific times that come to mind is one, I was driving and so I was sort of, you know, I needed to get something done, and I wasn't able to sit down and type it because I was. I was driving across the state. And so I dictated to it some thoughts. I was like, okay, I need. In this case, it was a letter. I'm like, I need a letter. Here are the points I'm trying to make, you know, And I just kind of. I spoke to it for about five minutes. It. And in this case, I actually used Copilot because I needed a Word document. So I did it through Copilot. So it was giving me a word output, and it gave me a pretty good first cut. It took my thoughts and put them in enough of a sentence structure that it was still as if I had drafted. And then, of course, I spent some time going through and editing it, but it was still largely my voice, and it had captured the things that I was trying to use. The other was a similar situation where we were putting out a report to our membership. And it was one of the things that, for me to write it, I was going to sit down and literally kind of look through our calendar over the last year and go, okay, September, we did this. October, we did this. November, we did this. And then I was going to have to organize it. And so I had Chet do that for me. So I literally. I just dictated my calendar. I did exactly that. I said, you know, month of September, we had this event, month of October, we put out this report on and on and on. And then I prompted it to, okay, now take. Take everything I told you and organize it into these categories. And so I, you know, I had an advocacy category, I had a professional development category. And it did. And it sort of gave me a first cut based on all the information that I was able to talk to it about. And then again, I spent a few hours going through and doing some wordsmithing, editing and moving stuff around. But that. That first, you know, that first draft that sometimes takes you a little while and you gotta be really focused. It, you know, it probably saved me a couple hours of work to just kind of put all that together for me in a way that I could start to manipulate.
- [11:32]
A
Yeah, I had a similar experience, and I think the dictation piece of it has really been the level up for me. And I have to give credit where it's due. A friend of sbi, Aziz Agaev, has really been a. Our AI genius on the podcast lately and across the association. But he and I have been talking a lot about the use of AI and He really opened my eyes to using the dictation function. And just a few weeks ago, I was driving into work and I was looking at my calendar. I'm like, oh my gosh, I have this goal setting meeting with my superintendent today. And I completely thought it was a different day. I mean, I had a rough structure of the three goals that I wanted, but I ended up just chatting in the car and my commute into work. And by the time I got to work, I had a fully fleshed out goals document that I had to tweak a little bit once I sat down on my computer. But essentially a 90% done document that I was able to bring in to my goal setting meeting with my superintendent. I'm like, this is kind of like jaw dropping for me because I was able to just literally have a conversation. And it is creepy, like Uncanny Valley ish, where when you're speaking to it and the voice is talking back to you, it just sounds like a buddy of yours. Which I was like, there are a couple of moments, I'm like, oh, wait, I'm still talking to a machine. Which is kind of bizarre. But I have to say, just being able to dictate in your car like you're talking on the phone, you know, because I have a longer commute, so it makes sense. But to be able to do that I just think is a game changer because I can accomplish something in that idle time. You know, it's just been, it's been a huge kind of level up for me in using AI.
- [13:12]
B
Yeah, absolutely.
- [13:14]
A
So what are some of the biggest lessons you've learned? But when we, when you think back when we were talking In January of 24 to now, both kind of is in its successes and limitations since really diving into this all in.
- [13:26]
B
Yeah, I think the, you know, the biggest thing continues to be for.
- [13:32]
A
For.
- [13:32]
B
Us and our work is making sure that whatever we're pulling it for is accurate and whether that's, you know, sort of that search function that it's providing. And you know, I never take if I'm used, other than if I'm just trying to. Sometimes if I'm using it to just inform myself on something, I'll. I'll sort of take what it says at face value. But if it's something that I'm trying to, I want to use in, you know, in some sort of product, I always do go back to the source document just to make sure that it's accurate. And you know, and it's still, it's still not 100%. So I think it's important that, that we do that. So the second thing that I've, I've also noticed it for is, is helping with sort of tone of voice. So an example of this is, you know, you, you may have seen. I recently wrote an article for our New York magazine that, that got published in In School Business now. And it was about, you know, sort of an ode to public service. It was titled In Praise of Public Service. But, you know, we've, we hear a lot of pressures on public school officials and you know, and I've been in government for a long time and so I just, I had some thoughts, I wanted to get down and you know, I think, I think I'm a pretty straight shooter when it comes to things like this. I'm, you know, I'm trying to be, trying to be sort of a nonpartisan middle of the road when I'm talking about this because I think these concepts apply across the spectrum. But I also recognize that we all have some inherent biases. And so when I wrote this, I did a pass through with ChatGPT and I said, and I sort of gave it my worldview. I said, knowing that I'm trying to achieve sort of a completely non partisan product here, and here are some of my own internal biases. Is there anything that I'm saying here that it seems like an inherent bias that somebody might take offense to? And it did give me like, that's pretty cool. Yeah. And it gave me like three sentences back that said, yeah, maybe these three you might want to rewrite now.
- [15:37]
A
Stop talking about fracking.
- [15:39]
B
Right. No, but, but it was, it was things that I thought I felt I had sufficiently wordsmithed it enough that I thought I was being middle of the road. And, and it helped sort of point out. Now I could have done that with a human editor, right. But I would have had to find the right audience that, you know, okay, you're the type of person that I want to make sure that I'm not offending when I'm saying this. Can you look at it right? How do I find that right people, all that stuff? So this was a way that, you know, I could, I could direct it to be that person for me. And sure enough, it, you know, it helped me clean some of that up.
- [16:09]
A
You know, I don't know why this popped into my head. I think just when I'm visualizing what you're doing, like asking for feedback and things like that, is this your experience? It doesn't seem to know when to stop it's like, well, I can do this next, and I can do this next. I wonder if there's a way you can just say, you know, after I ask you to do this, unless there's something really substantive to offer, like, don't keep offering, like, minor little things. Or, like, want me to convert this to PDF or do you want me to reword this? It's like, no, we're good.
- [16:35]
B
You know, it's funny you said you.
- [16:36]
A
Seem to be, like, incessant.
- [16:38]
B
Yeah, I was actually thinking exactly that. This weekend, in fact, you mentioned Aziz. We brought him in to do some staff training for the ASBA New York staff. And so one of the things he had kind of talked through was, and I know he talked through with your podcast, is sort of the creating your own GPT within the system targeted for certain tasks. And I was thinking this weekend, I have been kind of annoyed with that sort of endless, like, I'll just keep going. And I almost want to create a GPT within itself, that I only use that just to say, like, no, answer the one question I have and be done.
- [17:12]
A
Yeah, that's one kind of annoyance. And you're probably right. You could, if you use a GPT to be like, can you be a little bit more curt? That would be appreciated.
- [17:20]
B
Well, the current. And I don't need. I don't need the praise every time.
- [17:24]
A
Right.
- [17:24]
B
Like, oh, that's such a great idea, Ryan. You're so insightful.
- [17:33]
A
So I'd like to dive a little deeper into kind of how what we're talking about in terms of the use of AI is really benefiting Aspen, New York, and its members. So, you know, when we last spoke, you build a suite of different resources for members, from state aid tools to legislative data trackers. How are those tools being used and received by the members compared to when you first launched them?
- [17:54]
B
Yeah. So, you know, I think, broad strokes, I'm still largely using it to help me engineer within Excel tools. You know, this is, you know, we're using publicly available data, so we're not pulling anything from, you know, from district budgets or from. From our membership system. This is stuff that the state education department puts out on its website, and we're just building this within Excel. But I think the biggest sort of continuing improvement it's helped me with is just finding new and better ways within Excel and other tools to do these things. And, you know, I. I was pretty proficient in grad school and in my first job in using Excel, that was about 20 years ago now. And so there's a lot of habits I have in excel that using ChatGPT has made me realize, oh, Microsoft has made the product better over the last 20 years, and I don't have to keep doing it the same way that I did in 2005. And so that's been. In fact, I had never heard of the xlookup function before. Apparently it's now been around.
- [19:02]
A
That has unlocked new doors for me. Vlookup is a thing in the past, you know.
- [19:05]
B
Yes, no. And so, so I didn't even know it existed until in recent months because of ChatGPT, I've just been doing Vlookup. And so now that I know, like, oh my God, this is saving, you know, every time, every time I'm doing one of these functions, it's probably saving me five minutes in these.
- [19:19]
A
So what Vlookup should have been from the onset.
- [19:21]
B
Yes, yeah, absolutely. So, you know, that's where it's been been. Its biggest help is that, you know, and I'm sure, I'm sure there's, there's courses out there. I'm sure there's again, go searches, those kinds of things. But the ability to tailor it specifically to what you're trying to ask and get sort of immediate answers and feedback has been incredibly helpful. And it has started to make the charts that I've been putting together. It's really expanded how I'm able to make even things like the labels a little more dynamic than how they are. I would often be stuck in from creating a bar chart. There's only certain numbers I can use. And so I have to kind of build everything around. What am I limited to now? It's sort of really wide open that I can put the data in that I want to see. And then it's helping me figure out what are all the ways that I can manipulate sort of the display in order to tell the story that I'm trying to tell instead of having to reverse engineer a different story because of the data limitations.
- [20:30]
A
Yeah, that was always my experience and I consider myself pretty good with Excel, but there's always some real niche formulas that I don't use all that often. So a Google search would yield some really specific example, but not specific to me. So to your point, you have to reverse engineer what somebody else is doing to kind of make it fit your own. And now with ChatGPT, you can literally just upload the spreadsheet and be like, this is as far as I've gotten. This is where I want to go. And it'll kind of fill those gaps.
- [20:57]
B
Yeah, absolutely. And, you know, and oftentimes if, especially when I'm learning the new function, if, you know, maybe I've typed in the wrong sequence or something and I'll, you know, it's, it's so easy now to just put that formula in and say, I'm having trouble with this. And, you know, I'd say nine and a half times out of ten I'm getting, you know, getting the right answer back. Sometimes it's a little, you know, we're a little off on what, what it thinks I'm trying to do versus what I am trying to do. But yeah, it's been, I would say, I would say my Excel proficiency in the last six months has increased dramatically from what I'd probably hit a plateau 15 years ago and I've probably been at that same level and this has really helped me expand what I'm able to do.
- [21:37]
A
Think your pension is enough for retirement? Think again. National Life Group specializes in 403 and 457 supplemental retirement plans that can help supplement a pension gap. Their fixed indexed annuities are built with your plan in mind. Their indexed annuity policyholders take comfort in knowing they haven't lost a penny of their policy value due to market downturns and they can count on guaranteed lifetime income in retirement. It's time to secure your future. Check out National Life Group in your retirement plan listing or ask your HR team. Is National Life Group in my plan? If not, ask for National Life Group to be added. So what kind of feedback have you received from your members and how has it shaped what you plan on doing? Building next for just ease of data access and being able to extrapolate out what is otherwise very complex state aid data.
- [22:25]
B
Yeah, so our, as you know, and maybe some of your listeners have heard, we last year we went through a pretty extensive process in the state where the state government had tasked in law this research institute, which is part of the state university system, called the Rockefeller Institute of Government, to undertake a study of our foundation aid formula and come up with recommendations. So when we first spoke, I had done a first series of data tools that was trying to help explain foundation aid and some of our other formulas. Last year we ended up not really updating any of these just because there was all this uncertainty around the Rockefeller Report and it was published in December, the budget came out in January, so there was sort of, you know, just a lot of unknowns. And so now that, now that we've gone through that process, that the Government did make some changes to the formula. We're now getting back into a new budget cycle where I'm going to be able to sort of reproduce those with some of these new capabilities. And so one of the things that we've done related to that is we've built a professional development series. It's six, we did six hour long webinar sessions over a six week period. But as part of that webinar, which is largely just hearing from a speaker and being able to ask questions, we're also providing some new updated tools. And so we're waiting to get this feedback because folks haven't gotten these yet. But in, in this world where we've now been able to do a lot more with this public data and how it shows up in an Excel report, we're going to be sharing this out as a part of this workshop series. And I think this is potentially the starting point for doing more of these because once you have the body of work built, you can just update the data and continue to expand them. So that's where I'm really excited about the possibilities we have. Obviously we're still waiting for feedback, but at least in the folks that I've beta tested some of these with, the feedback has been incredibly positive. And quite frankly, the ones that we've shared to date are still, in my view, pretty basic. I think we've got some really exciting versions that people are going to really like once we were able to publish them.
- [24:56]
A
So you've been able to take just large swaths of state data and really pare it down to an individual district level, is that correct? Are you able to now make it a little bit more meaningful for an individual to district so they can maybe select their district from a drop down, like, oh, okay, so this is how the state is or how the state data is impacting me now. And what I can anticipate kind of moving into the future is that kind of how this has been like developed over the past year or so. Yeah.
- [25:24]
B
So the way our state data works is there's a large Excel file with multiple tabs. It's based on a 40 plus year old mainframe system. And so it's not like one big database that you get, you get this spreadsheet that has multiple tabs and all sorts of crazy code. So you have to reorganize all of that. But at the end of the day it is district level data that is shown in the aggregate. And so what I've been able to do in sort of wedding my personal knowledge and experience in this world with the capacity that Excel has with chat, sort of being the mediator between those, I've been able to find ways to take that individual district data and out of the hundreds of or dozens of data points, really hone in on what are the two or three important ones in this particular area and then find a way to graphically represent that. And that's another area Too where using ChatGPT has been helpful because I will often have, I'll have an idea of how I want to display this data and sometimes it doesn't quite work out and so I'll take a screenshot of it and say, all right, here's what I'm trying to. Here's the data that I'm looking at, here's what I'm trying to express in it. This is the graph that I have. I don't think it works. What other ideas are out there? And so similar again to this sort of replacing search function, it's really able to give me answers specific to what I'm trying to do and not sort of what are the menu of all the things. And then I have to try to figure out what's the right one for me to pick.
- [27:10]
A
Right. So you'd mentioned earlier, it's not always accurate and you have to use your own professional judgment and proofreading, obviously. So how do you balance the innovation of all this with the responsibility of accuracy trust, especially when using AI for public facing tools like we've been talking about today.
- [27:29]
B
Yeah. So, you know, actually in preparing for this conversation, I was thinking about how do we, you know, how do we think about this? And I actually came up with sort of three, three different models for thinking about it. I think they're all trying to get to the same place. But the first is that focusing on content creation versus curation. And so like I said, we try as much as possible when we're designing something, writing something, if we're using AI in any way that it's not being the original content creator, it's really helping us curate what we have in coming to that. And I think that helps with that input process. Right. If we know that the inputs are not 100% accurate, even if they are very accurate, the more that we are controlling what's going in, the more likely that's going to be correct. The second is thinking of that we're replacing process steps, not the products. And I think especially when you're thinking of, you know, smaller organizations, maybe if you don't have a lot of staff you, you may often, you know, it's easy to, to put out flyers and things using ChatGPT. And I think, you know, when you're looking on social media these days especially I see a lot now with, with like restaurants and bars. So if you, you know, if you follow, follow businesses on Facebook, they, you know, there'll be, you know, Friday trivia night. But you can, you can tell that like you've seen enough of these, you can tell if it's an AI generated flyer, which is totally fine. There's no reason that people shouldn't use that. But if from sort of our brand perspective as an association, we would want to make sure that it is more in keeping with what we would do. And so thinking about we're not using this to replace a product that we're doing, but what are the steps along the way that maybe there's some, we're losing some time with some manual work that maybe AI can help us with. I think those are, that can help straddle that goal of making sure that it's accurate. And so related to that, and I think this speaks particularly to ASBO members across the country, is that this is supplementing but not supplanting our work. And again, there might be some ways that within process steps that it's supplanting, but a lot of the work that needs to be done at the end of the day still has to come from, from people who are doing the analysis and having the connections between the community and their administration. And there's a lot sort of under the hood that can be automated. And so thinking exactly about that, what are the ways that we're supplementing what we're doing, but not necessarily supplanting the core work. We can supplant some of these process steps, but I think as this has become much more ingrained in what we do, you can just see it, right, I mentioned the flyers and things you see on social media, but even in presentations and things now, right, you go to, you can just, you can sort of tell when somebody has used AI to generate their slide deck versus when they haven't. And I do think there's over time there probably is going to be, you know, a sort of a premium placed on when you can tell it's sort of original work, right? Because I think the AI generated products will become so ubiquitous that you will sort of stand out like, oh, this, this didn't come from an AI. And so I think, I think it's important for all of us to think about that. You know, that there are ways that we want to, you know, we want to still have that premium placed on the work that we do. And so just being mindful of, of, of how we navigate that, it's a.
- [31:21]
A
Great lead into to my next question. So what role does human judgment still play in your data work and advocacy work?
- [31:28]
B
Yeah, so for us, it is still 100% driven by the human element. As I said, that search function. Now, having the sources has been helpful. There definitely have been times in our advocacy work that I don't necessarily have all of the subscriptions to some of the legal services that if I did, I would know right where to find the answer, but in the absence of it trying to find one. And so actually this was a use case I used over the weekend that if I had a Westlaw subscription, I could very easily look up the legislative history of a particular section of law, know when it was created and all that. But because I don't have that, I was able to. I asked ChatGPT, you know, this particular section of law, and it did a little bit of searching for me and, and I'm 90% certain it's accurate now. I wasn't using it for something that I'm publishing in anyway, so I'm fine with the 10% uncertainty I have with it, but that's a place where.
- [32:36]
A
I.
- [32:37]
B
Could have just copied and pasted what ChatGPT gave me. But because I don't have that 100% certainty, I'm not totally prepared for that. So I think just the, the making sure that being both the originator of the idea, but then ultimately owning everything that comes out of it is sort of where we've used it. And so there are, you know, I think there's potential use cases that we have chosen not to use it for just because we couldn't feel comfortable at both ends of that process, both at we were complete owners of what we were putting in and that we could completely verify what we were putting out. When I say putting in, I mean we have still as an association, we have not used any of our membership data or anything of that nature in an AI use case. It's really just been more on some of the communications, tweaking around the edges sort of thing and obviously the, the tools that we've been talking about, but we've been very mindful to, if we don't have complete comfort with what could happen to the information we're putting in, then we don't want to put it in.
- [33:55]
A
Right. Well, what about the rest of your staff at ASBO ny, I know you're all in on using AI. Has the rest of your team used this technology to kind of change how you work and collaborate internally?
- [34:07]
B
I think there's been different degrees of it. I know with respect to our website, we've had some design issues with our website that we haven't loved over time and we don't have a company or a full time developer with us who's done that. So I know staff have used AI to help make some modifications there. And I think for someone who, if you have folks doing this kind of work and they're not computer programmers, maybe, you know, maybe the, the coding language isn't intuitive to them, but. Sure, but AI can sort of walk you through that and you know, and we've seen just some of the improvements we've seen in that. I do know there are some other, other use cases around again, sort of, you know, more, more in the editing and communications place. But I think, you know, as a, as an organization we've, I think we've had a healthy level of skepticism about it and each of us are a little bit different. I've probably, I probably use it the most. But again, a lot of my use, a lot of my day to day use really is just in my own personal life and just sort of navigating, you know, navigating the day. I have a, I have a question, I have a thought, you know, something that maybe it'll take me a while to Google. I helped my son with his homework the other day and it's been a long time since I've done sixth grade math. And so I'm looking at a question like, boy, like I know the answer but I don't know what the teacher's looking for right in terms of the work. And so that was when I, you know, I snapped a picture of it and I said to ChatGPT, you know what, like what's the, what's the right process here for 6th grade math? And it gave me, and it was, and this is actually a good human validation point because my son was struggling with it and, and I was struggling because I didn't know what exactly he needed to be able to show. And so, so I pulled this up on ChatGPT and I said, oh, you know, is, is the answer. You need to put these things together. And then the light bulb went off for him because he knew it, but he had just forgotten it in the moment. So once he sort of validated for me, yes, that's how I need to do it. And then he blew through the rest of the questions and he was able to get the assignment done. But that was another important place where here even an 11 year old is helping to validate what it's putting out with his own knowledge.
- [36:26]
A
And I think that's going to be important thing moving forward, at least from the student experience is I see AI being a huge help in terms of like a tutor and getting your mind unstuck on something, not just taking a picture of a problem, pasting it in there and then copying the answer. I mean that you're not going to learn anything that way. But you know, we've done some, we've had some conversations with students at my district and they're all using it that way. I mean, they say, you know, every now and then someone's just blatantly copying, pasting it. But it really has become an incredible tool to kind of unstick their mind in terms of if they're, if they're stuck on a problem or they need a little more inspiration or you know, some writer's block. It's just been a great tool to, to kind of really supplement those, those skills, you know.
- [37:09]
B
Yeah. And you know, and I, that is probably my biggest concern with, with all of this is that when you and I are using it, we're using it with a base of knowledge that we already have and we're able to, you know, think critically about, okay, this works. No, this doesn't work. And so I really do worry that for, you know, for, for kids and young adults who, who haven't already developed that knowledge, if they're going to overuse it in a way that sort of removes that ability to think for themselves later on. Right. You and I can look at these things critically. I, I, I, I worry that, that they, you know, that they can't when, when the time comes, you know, and, and another good example just of, of personal use, I, I ad struggle with particularly classical fiction. I read quite a bit. But when it comes to novels that are written more than 50 years ago, there's a thought process that goes into the writing that whether I'm too distracted or whatever it is, I always struggle with reading comprehension, with that. And so I recently started reading James Fenimore Cooper's Leather Stocking Tales. I live near Cooperstown and I've heard about these forever. And here I am in my 40s and I've never read, read them. And this was actually for the first time I used AI in that Some of these stories are well Known there's no spoilers to me at this point, but I asked it for a couple bullets for each chapter, and I said, spoiler free. Give me a summary of a couple sentences of what happens in the chapter. So that now, before I read the chapter, I read these three bullets, and then I read the chapter. And. And it actually helps me with sort of whatever is going on in my brain that I can't process it. It's been helping me ground in like, okay, this is where we are in the book right now. And so now I can kind of focus on what's happening as opposed to feeling like I'm constantly lost in the language.
- [39:08]
A
So it kind of primes you for what you can expect so you're a little bit more engaged with the actual text itself rather than trying to figure out where you are in the reading altogether.
- [39:16]
B
Yes. Yeah. And, you know, and I don't need the CliffsNotes here. Right. I'm not trying to write essays or anything about it. I just. I need somebody who's read the book to tell me what I'm about to know. Yeah. Yeah. And so. And you know, and even then, you know, there's still. I think part of this is just in our day and age, the, you know, all of our attention spans are so minimal, and especially when you got two kids running around, Right. Trying to read something, especially that, you know, if, you know, the Cooper books are written in the 1840s and 50s. So just the English, you know, what we say now in 10 words, the sentence he's writing in 30 words. Right. And so there's just like a. There's almost. Even though it's English. Right. There is kind of a translation. Yeah. And so. So it's. It's. It's helping me sort through that.
- [40:04]
A
Yeah. I. I think I mentioned in my. I had a reflection episode last week, and that's one thing that has been terribly interrupted by having a second child is my free time to read. So I. I may need to take some of your advice and leverage AI a little bit to kind of jump start and speed up my. My reading consumption because I just. I can't find the time to sit down.
- [40:26]
B
I was gonna say my, My. My only advice, and I don't know that it's good advice, is if. If you have two kids and you want to get a lot of reading done.
- [40:33]
A
Don't have a third.
- [40:34]
B
Well, no, no. You need to. You need to stay up till one in the morning and only get five hours of sleep, and that's, you know, you can get some.
- [40:40]
A
That's a non starter for me, Brian. I'd rather just not read.
- [40:43]
B
Yeah, well, that's probably a problem on my end that I probably should be. Should be shooing the reading and getting some more sleep. But, you know, here we are.
- [40:52]
A
Oh, five hours. I don't even want to talk about that. All right, so when you look ahead into the future here, you know, the next year, what's kind of the next big project or vision you have for asthma, New York, kind of in this AI tech adjacent space.
- [41:08]
B
So, you know, this really stems from, you know, we have a strategic plan in our board which you're a part of, has just refreshed it. But in addition to our official strategic plan, I have always had what's called a big hairy, audacious goal. For those who aren't familiar, this is Jim Bhag, Jim Collins book, Built to Last. Great, great book on management, if you're interested. But the idea behind a BHAG is that it is a significant goal that is still actually obtainable. Right. This isn't some platonic ideal that you're never going to reach, but you want to keep working towards it is something that with the right amount of work you can put in and get there. And so my own personal BHAG for the, for the work that we've done is that so in New York State, school district budgets are voted on by their local taxpayers. And so my BHAG here is that every voting taxpayer in New York State is better informed of their school budget because of the work that ASBO has done. Now that's. That's coming from any resources that we put out. That's coming from media hits that we do as an association. That's coming from the trainings that we do for our members and helping them to put together and communicate their budgets. That's the resources that we put together for them to report on these for their folks. And so I think everything that we do can really distill into that. So as I'm looking to the future, how can we leverage these tools to help make that better? Right. And I think the uses that we've had for some of these Excel reports is a really exciting first step into that to expand out both what we can give school business officials to use in their work, but then also what can we be putting out there publicly just for the general public to have in a way that supplements what they're doing, what our members are doing, I should say, without necessarily stepping on toes or interfering with sort of the local budget development process. And I think that's really going to, there's going to be a common thread there throughout, not just the resources we put up, but also how do we develop professional development? What are the topics that we want to cover? What are the ways that we can be communicating this to people? I think it helps sort of open up the possibilities that we have.
- [43:38]
A
Yeah. Well, Brian, it's, it's amazing to see you kind of progress throughout this space and just to think about what we were talking about, you know, almost two years ago and just kind of scratching the surface with ChatGPT to now being expeditious in Disneyland, you know, it's pretty impressive. So I'm excited to see what you could do in another 12 months when hopefully I can have you back on and we can reflect even more on what you've been able to do. And on a personal you've done such an amazing job with your staff with ASBA New York, and it really has nothing to do with AI, but just your leadership and the staff that you've put together. So from me to you, thank you and thanks for coming back on School Business Insider. I'm sure you'll be back on in short order.
- [44:20]
B
Thanks. Yeah, no, I appreciate the kind words and always happy to come back.
- [44:26]
A
Thank you for tuning in to School Business Insider. Make sure to check back each week for your favorite topics on School Business.