Information Technology Advisory Committee - Regular Meeting

Alright. It is 06:00.

So it is. Alright. I will call the 11/05/2025 meeting of the information technology board to order. All members of the committee are present. So we do have a quorum. Item two on our agenda is the pledge of allegiance. I guess I'll have to my left.

It's like the

liberty to the flag of

The United States Of America and to the republic for which it stands, one nation under God, indivisible, with liberty and justice for all.

Item three is public comment. There are no members of the public here with us this evening, just staff and the board. So we will move past that. Item four is approval of minutes. We discovered before the meeting that the minutes that are attached are from our June meeting. Those were approved at our last meeting. So we're we're meeting behind in minutes. We should have the September minutes, but we do not have those tonight. So we're gonna move past that, and Chris will work on getting those attached to our next agenda for approval. That takes us to item five, which is our business items tonight.

Our first item five a is to review and act on the amendment to the agreement with Central Square Technologies to include the law records management system. I'm gonna turn it over to Chris, Greg. Yep. Yeah.

So chief Thompson from the police department's here joining us this evening for this item in particular. So back in June is when we brought the the CAD system before ITB that subsequent subsequently went on to common council. It was approved there. We started doing our our project meetings just a little bit with Central Square, starting to ramp up to get in the full project mode with them. While we were doing that, we were in discussions with a a different RMS vendor.

And initially, when we are looking at who to do the records management system with, this other vendor, while a little bit more costly upfront than Central Square, Year over year, was, was less than Central Square. That vendor was Axon, and so police department uses Axon. That's the they own taser, body worn cameras, fleet, interview rooms, all that stuff is already Axon. And so it made a lot of sense to try and have that complete Axon ecosystem when it came to the records management part of it. But as we progressed, price creep set in with it, and then they eventually priced themselves out outside of our budget.

So went back to Central Square. And last November, they had given us a a best and final offer for an RMS system. And I reached out to our account exec and said, hey. Would you honor this this baffle? And he said, yeah.

Let me take it to my my boss and see. Came back a couple days later and said, We we'll honor that. So we went through the the process of the redlining the scope of work, which is which is an amendment to the CAD scope of work, whichever one was already approved. So, you know, there's additional interfaces that come with the records management system. But there was we initially, we had an interface from Central Square's CAD to a third party or a second RMS vendor.

So that that interface will come off of the overall project cost eventually through change order. But so what we have before us is and what was attached to the agenda is the statement of work, the amended statement of work to include all the law records management stuff. So just to, make that distinction, the fire department already has a records management system. So this is a law records management system. And, if you have any questions, chief Thompson's here, and I'm here as well to answer them.

Chief, did you wanna anything you wanted to say as a primer?

Well, I'll just add that. Obviously, the CAD period dispatch and the RMS is like the light blood to keeping our public and officers safe. I will say this, I personally looked at Central Square's RMS, and I was actually impressed. The reason why I say that is my prior life in Milwaukee. We had Tri Tech and Timuran, which was products that were bought by Central Square.

Okay.

And we were getting it's like a Tri Tech version, an enterprise version. I was kinda cautious because what I remembered and all the problems that we had in the past, when I met and actually looked and actually got to work in this new environment, it's night and day. What it's gonna do is make our investigations more efficient. It's gonna make all of it more efficient and actually help our officers be on the street more than being inside doing reports and things of that nature. So I'm really looking forward to getting this going. In fact, I wish we could

get it

going tomorrow. Yesterday. Yeah. Okay. Okay. Go ahead, Steve.

So chief Yes. Used to work for a company that had something like this record management system and it it would work with the DA's office. It worked with the AG, state AG and it also work interface seamlessly with different agencies in the federal level. Is that you know, you were alluding So it's pretty seamless. So we can This will be

more city centric. Mhmm. But there are some other technologies that we have already existing within the Waukesha Police Department that talks with our DA's office. Mhmm. Now in the future, we'll be coming with some other technologies to information sharing technologies, law enforcement centric.

That will specifically address what you're talking about there, where you have, in the state, you have you can count three different CAT RMS systems in the state, and they're all operated as independent silos. Our goal is to try and get, like, an epic chart for for law enforcement, and we're working on that, and I foresee that in the future. Right now, this CAS RMS is it moves primarily for the city. It's gonna really elevate our ability to share. It might not be real time internally through digital, I should say. But the information we'll we'll be able to produce, we can actually share via email and other other methods.

We do a follow-up, Dan.

Yeah. Go ahead.

So chief, when when your officers are writing their reports, what what time percentage of their time would be saved with the system versus, you know, handwriting? It's quite

So with one of the what do wanna call modules over here?

The centerline.

Centerline. What that's gonna do is I could actually see and reduce our time by I'm being very conservative. See it's more than 50%, more than 60%. Okay. Because what I can right now, even without this future software that we're talking about bringing in, just with this technology itself, the central square SquareCAD RMS, its capabilities, capacity, and how quickly it can bring data to the forefront. It's just gonna save time by itself. Now you add that to report writing, I think it's gonna be a big game changer for us.

Yeah. The quality of those reports too with the with this product because it's making sure it's checking all the boxes for Not

just quality of the actual report, but also the quality of how we do our interviews. It'll actually teach the officer to interview the right way because what I'll do is I'll take that front camera and take a transcription put onto a report, but then it forces that officer to actually go through it and make sure that everything's accurate. And if it's not, I'll actually put almost like an alert. Like, hey. It looks like you missed this. Looks like you missed that. Looks like you missed this. And I'll say historically, I'm at the dinosaur age of my career. Right? So when I first started, I'll give you a little story.

And I'm doing homicide interviews, and we do it on paper. We intentionally make a mistake on the report so the suspect could actually initial that that was a mistake and write what the actual sentence should be. What that did is show that the suspect actually read the whole thing and that the report was

accurate. Interesting.

So then they told us we have to go to recordings. We were all complete.

Like, why do we have

to record this? And we didn't wanna change. But when we went to recordings, it was the best thing ever. Yeah. Well, then we were told, now you have to videotape them. We were so resistant to that, but we started doing that as the best thing ever. See if it's gonna happen here is this is the future. How we leverage AI, how we leverage technologies is is the future, but we've to be fiscally smart. We also got to be responsible for making sure we're following all the constitution and privacy and those type of things. But we also have to leverage every tool and toolbox to keep our officers and our citizens safe.

Any other questions? I'm ready.

Still have Chris, you mentioned that we're, I guess, eliminating the interface piece that we were expecting to have between Axon and Central Square. Don't we still have an interface between Axon because they're driving off of our body cameras, or is that built in already?

It's the Axon RMS interface that will go away. The Axon to the Central Square to axonforevidence.com, which is how they share that evidence with the DA, that is that is an interface that's included with this. Okay.

So And they both will talk to each other.

Yeah. Great. Great. Yeah. And is it DA Protect? Is that kind of what Steve's talking about there? What how DA then you can share information with the DA. There's a product called DA Protect. I think our the DA just bought

it. Correct.

Yeah. So

Do you have to wait for the whole implementation of everything before you can get this going, or is are they hooked together?

Or No. We we're gonna do them we're gonna do them side at the same time. Because we're gonna do a big bang cutover.

Oh, mhmm.

And so it's a this is probably a fourteen to eighteen month project before we go live. And so there's well, there's just a lot to configuring this system. Everything that we do now with the current system is being scrapped because the current state is not what we want for our future state. And so really, like, some of the vision that chief Thompson has with, intelligent led, data driven decision making. That was perfect. Been hanging around with him a

little bit.

But those initiatives like that he has there are gonna be made possible with this. And what this system is capable of doing compared to our current system is night and day. So our current system, if you're gonna tier these different systems out, it's a tier three at best. This is a tier one system. And and what it what it can do is is is pretty incredible.

So as far as how they capture the data, how how it's the quality assurance process throughout everything is is is really good. So so, yes, we are doing them both at the same time. And so we've we've kinda pumped the brakes on the on the CAD waiting for this to get caught up, which is good because it it's allowed us internally to get all of our ducks in a row. And I'll touch a little bit more on that with the next item that's on the agenda. So we don't wanna keep him waiting. Oh, good. No. No. Are are we

still on the same timeline or did we get If we off a little bit.

If we're off, we're off by a month or something like that. But it's all but it's to our benefit now too because as we started going and working with our consultant, we've just realized, like, the amount of subject matter experts that we need involved, the size of our core team, and just kind of our overall project governance and structure as as it looks a lot different than it did three weeks ago today. Okay. So so we're ready for it now. And so we're this is good because it's gonna it's given us a month now to get things like our strategic objectives, for this new system down on paper.

So we have something measurable to to look at, you know, a week, a month, six months, a year after our go live. Are we hitting these objectives? And it's not just police objectives, fire objectives, but it's joint public safety objectives as well. So there's there's all sorts of things that's given my team a chance to build the infrastructure. So we 45 servers altogether between the CAD and RMS.

So, you know, six of those are, like, SQL servers that are in an always on cluster and stuff like that. So some of the cool stuff, the way Central Square does their database is they only keep about six months of data in the live, database, and then everything's pushed out to a read only database. And so, like, anything older than six months, you know, when you're doing a query that's having to go pull it from from that other database that we're And then additionally, all of the interfaces touch that read only database. They don't touch production. So it's like built in our

I love that. Yep. That doesn't touch production.

Yeah. So it's they don't call them data lakes. They call it a data pond just because it's it's not super huge. So, anyways, and then all of our historical data from, Phoenix will will go into that live, read only database, so it'll be there. But then we even have, data from the a s 400 system and put that's been converted to SQL format.

That will go into an archive. And, but when when they're deep querying data, not only does it do the read only database, but it'll go out to this archive database and pull those records that were out there from, you know, the nineties and systems. That sort of thing. So

I got I got another question With for that system, you know, you're talking I think we talked about this a couple months ago that you were kinda shy of, you know, servers and that sort of thing for a test bed. Do you have enough equipment to do a proper test bed on all this? And then Yeah. We had phase in with the police department so that some officers, detectives, lieutenants, they can get on this thing and and give you feedback before you actually do the full cut over.

Yeah. That's all part of it. We stress test that thing. We did have to buy an extra shelf for our Nutanix setup Mhmm. Because of because of the amount of servers.

But, yeah, this the product comes with production, test, and training and and Doctor. So it's fully redundant. Disaster recovery is built into the product, all that stuff. And Central Square, once everything is set up, they do have a program that runs to load test it. So we can simulate all kinds of different scenarios, and it'll run, and then we can see how the system's reacting to it. So yeah. And then and then there will be the the parts or the times of during the project where we are specifically trying to break it.

Where'd we come in in terms of budget, I don't remember what this you said 3,000,000 that this project Yeah. Originally

CIP? Yep. The the project was 3,000,000. We're still sitting at a really healthy number from a budget standpoint, which is, we have, I don't know, somewhere 20 some odd interfaces. Some of those companies will charge us to develop the interface for Central Square or to give us the interface for Central Square. Some of them, might be a, you know, 5,000. Some of them might be 10. We know one will be a thousand. Some are free. So that project money will be used for for stuff like that to make sure all the you know?

Because those are the one things that we don't know yet. Sure. So we just started reaching out to all those different companies to to get those, signal We're

not expecting to go over that 3,000,000 number. No. No. Okay. We're No. We're in a

really good spot from a

I know that's been I know that's been a concern throughout this project at different points. Well, yeah. You know, when you that on the record.

When you budget for something three years ago Right. You know, and and and there's a lot of just unknowns going into it. Right. You just hope that it was a good number, and and it was. So

Any question? Go ahead.

Out of curiosity, so you said six months of data. Anything older than that goes into read only read only database? Yep. Is there any other sort of, like, cold storage, or is it, or or what what happens to data that's does it is it sit there does it sit there indefinitely then? Yes.

Yes. Yeah. There's database. We don't nothing comes nothing's ever purged from the database.

Right. But what about backups of that?

Yeah. Those backups just get bigger. Okay. Yep. Okay. I'm curious.

Yeah. But it's indefinite storage, essentially.

Yep. It is. And that's how like, our current database is somewhere six to 800 gigs in size. You know? So when we back it up, dispatch knows it. You know? It's just it's such a long backup that, you know, it's hard not to feel some kind of performance hit. Yeah. Like, they're you know, that stuff, you just can't you can never purge it. And, you know, one of the big reasons why we're bringing all the data from the old system into there instead of going into that archive tool is because there's open cases.

Sure. You know? And some of those open cases might be ten years old. You know? And so you you don't wanna miss not, you know, not bringing in something like that. And, you know, there's a bunch of other, really good reasons for doing it. So

Some of that stuff, I think, by records retention law has to be kept essentially indefinitely, if I remember.

Right? Well, yeah. And quite frankly, it's just all useful data. Right. Right. Even if the quality of it isn't super good, you know, from where it started out to where it is now. But all that data's got used to it.

With advances of DNA continuously progressing, You wanna be able to maintain full sources because when we cannot connect the suspect to today, tomorrow could be a different story.

Sure. Sure. Is there a

question, Chris? Steve, remind me citywide, how how big a pipe do we have running everywhere?

Well, we have our own private fiber fiber network everywhere. So it just depends on the optics that we have in our switches. We're a 100 gig to our our disaster recovery site. And, you know, I think the slowest connection we have is 10 gig.

And that's to the pools or

something like Yeah. Okay. Yeah. So we can We're plenty fast. We can have

a range of redundancy depending on how we wanna

We have we have primary routes, secondary routes, a lot of our places, even tertiary routes. So we got a pretty good mesh network.

Last Thanks. Chance, any other questions, comments, concerns? Think we're looking for a motion to approve.

We do have suggested motion language here, and I want us to take a crack at reading it.

Okay. I will make a motion to approve the amendment to the agreement with Central Square Tech. Acknowledging is to include the law records management system module in the amount of $926,005.69 and 52¢. I have a second to that motion. I'll it. I heard Steve first. So it's a motion by myself, seconded by Alderman Bantrieste. Any other questions or comments before we take a vote? All in favor, signify by saying aye. Aye.

Any opposed? Any abstentions? I believe that passed unanimously.

Awesome. Thank you. Yeah.

Thanks for

coming in. Yeah. You too. Thank you.

But we have Tom's contract next

if

you wanna hang

around. I'd love to. Won't. Yeah.

Somebody. What's that? You're missing somebody. There's seven of us. Looks like Paul? Paul.

Oh, man.

Hey. Mister mister mister Favy and our vice chair.

Interesting.

That'll be corrected in post production. It will. Yeah. Will. Okay. Moving on item five b, review and act on the oversight project management scope of work with Winborn Consulting for the CAT RMS project. So this is a connected item to the one we just talked about.

Chris? So we've been using Tom from Winborn Consulting from the very beginning of this project. I met Tom. You were introduced to Tom during the CAD. He came in virtually.

And and just looking at, you know, moving forward with this project, we want to make sure we still had someone with the expertise that Tom has. You can I got just a handful of some of his recent projects up on the screen there? And the ones highlighted in yellow are ones that he that are Central square projects. So Windborn Consulting does not you know, they want their customers to find the best possible solution for for them, and they manage. But they've done the city of Milwaukee, that's one Chief Thompson's familiar with.

That's a product called Hexagon. So Tom's been working with Milwaukee on that. City of Virginia Beach, I believe is an Axon project. And then but the ones highlighted in yellow are Central Square. So he actually has a current project going on with a customer that's a Central Square project.

And then and then they are close to another contract that's also a Central Square project. So Tom's got a ton of experience doing this stuff, and, he's super helpful. He's he's always given us the ability to just use them a la carte. So, but what we what we wanted, him for was oversight project management on this. So what he did is he gave us a a scope of work for project management oversight.

That is really just attending a lot of meetings. It's I'm the I'm the lead project manager on it, and so he's kind of really consulting me on making sure I'm I'm I'm doing what needs to get done, giving us a lot of good advice as, making sure we're aware of all the different gotchas on a project this large, and and that sort of thing. So if you look, at the at his scope of work, you'll see it's roughly a $120,000 for the project oversight. And then a little bit below that, there is a a list of other items that that he can help with and that he recommends that we do. And so that is where the the not to exceed $238,000 comes from.

It's a $120,900 for the base project management oversight firm, and it's detailed out in that scope of work. And then like I said, if we chose to go with every other item that he has listed in that scope of work. That's what the total would be. So

Was the intention was the intention all the time to have him or their company provide consulting throughout? Or We I I I guess I'm I'm asking because, you know, he's already been involved. Yeah. Like I said, we we met him virtually virtually as part of the presentation of the initial CAD system. So I guess I'm just kind of trying to understand, like, what was his role supposed to be and what is it changing to? Yeah. Or or is it changing or was this always the plan? I guess is my question. This became the plan.

So as we started using Tom more and more, it just be I mean, his his expertise and has just proven invaluable. I mean, when we look at just like these statement of works from Central Square and all these other vendors, you know, for their CAD and RMS. When you look through there, if for us, we read through there. It's like, yeah. This looks pretty good.

And then Tom reads through there. He's like, no. You gotta do this, this, this, and that. And it's like, you know, you don't you don't have someone like that that's been down this road before, and you just you're gonna end up getting to a spot where this system is now functional, and it's gonna get the job done. But all the other bells and whistles that we want will never get there. And so where Tom, I guess we never we didn't know that we needed a consultant until we started working with one. And I know, right, consultants are salesman. Right? They they do have to sell themselves, but Tom has not only sold himself. He's he's proven himself.

And and we just don't think that without him, this is gonna be a successful project.

Yeah. And if I can add.

Yeah. Yeah.

What he does is great job of is being able to translate vendor language to public safety and city IT language, and that is critical. I was at Milwaukee when Tom was there in Milwaukee, and I've seen when you have where you would do a great job, but the city didn't listen to the job. And they thought they could do a lot of it on their own, and they paid for it. We cannot make that mistake. So what you're doing and what we're doing here with this project because you think about it, this is a legacy. I just gonna last us a long time. We gotta do it right.

Yeah.

So this is well worth the investment. And even though he might be annoying at times because he's redundant, we're looking it up. That's actually a good sign. That's a good sign because he does not miss a thing. Yeah. He's super detail oriented, and we we need that.

So Yeah. Okay.

And, again, from a budget standpoint, we're okay with Oh, yeah. Where we're at with this this cost as well. Yes. Okay. Any other questions about this? Yeah. I I mean, again, we met him virtually, and he seemed very knowledgeable. His resume kinda speaks for itself. Yeah. But so I'm you know, I I just wanted to understand, you know, kind of why we were, I guess, adding this on or, you know, what the plan was from the beginning and what the plan is going forward. And Yeah. We've like I said onboard.

I just We've been using him a la carte. Yeah. And and now it's like, alright. Well, we're not just gonna keep a la carte in this thing and not really understanding what the full or what the potential costs are gonna be with him. Mhmm.

So give us the scope of work, and you can see he gave us he gave us four options. Well, the fourth is really no option. But, you know, it's full project management, which was really expensive. Oversight project management, which is kinda like co PM with me, and then project management block of hours. And that's where I mean, we could go buy a block of a hundred hours, use them up, and then go right back to them. And and he would just keep giving us the same level of service we're doing now. But this one here, we commit to him, and then he's gonna make sure he's gonna have the the room for

us too. So it's essentially, like, we're we're putting him on a retainer, if you will. Mhmm. Yeah. Is that safe to say?

Sure. Yeah.

As long as we're not dependent. Any

other questions, concerns, comments about assignment?

I'll make a motion to approve the oversight project management scope of work within onboarding consulting project. That amount not to exceed $238,120.00, including base services and optional task as selected by city staff. What word did Sean? Okay.

We have a motion. Do I have a second? Bob raised his hand. Is that a second, Bob? Yes. Okay. Alright. We have a motion in a second. Any other comments or questions before the votes? Motion by mister Anson, seconded by mister Brewers.

All in favor of approval signify by saying aye. Aye. Any opposed? Any abstentions? I believe passed unanimously. Alright.

Thanks, Steve. Appreciate it. Thank you, Steve. Thank you.

Again, Paul at post production.

Yep. I'm sorry. Didn't raise my head. That's

so high.

Can I jump?

Paul's here. We we promise. Let's see him. Okay. Alright. And then one takes us to Item 5 c. Review and discuss the AI acceptable use policy. I think we've I think we previewed this or talked about the fact that we were gonna do one of these

Yes. Sweet. Last year

maybe or early this year.

So I've been threatening. See it

coming before us. It's a popular topic. Right?

Yeah. Skynet

is coming. Right? Yeah.

It's already here. Hey, God help us very much. Alright. So I've I worked on this draft. City attorney took a look at it. A lot of this is coming out of, like, recommendations from NIST, like, ISC squared, ISACA, and so this is kind of that mishmash. And, again, it's a it's an acceptable use policy. So what we wanna do is we wanna allow our users to use AI where it's beneficial. You know, everyone's dipping their toes in the water to try and figure out how it can help them. But we wanna make sure they're doing it right.

And so part of this does in include, before being approved to use, AI is actually getting some AI training. And so in did we in the last, know before, well, it's actually this this quarter's know before. It touched on our security awareness training. It touched on AI a little bit. But know before does have a real kinda broad.

Here's what AI is. Here's some, you know, general uses. And then they and then they have some more, like, you know, they're short videos. So it's a slightly deeper dive. The UW extension is working on an AI training for state and local government.

Mhmm. Right now, they're hoping that that's gonna be available by next by next fall. So we're we're keeping our eye on on that. So then we'll have some good kind of formal training to put people through. So with that, you know, we're just looking at, you know, we go down through here.

We're just defining AI. We're defining what con confidential information is, what devices are, IT assets, all that good stuff. The, I guess, kind of boilerplate stuff here, and then we just have the rules of AI use. And then if we go down a little bit further, we have use of what happens with use of non approved AI tools, how do you get exceptions, and that sort of thing. So along with the rules of AI use is is also comes with the shall nots.

And, you know, the big one here is really confidential information. Like, we don't you know, you obviously don't want HR experimenting with AI and throwing everyone's PII out there. Right? Because that once it's out there, you don't get it back. It's just stored in a

I'm assuming we're using Copilot because of our integration with Microsoft.

The free version.

Okay. Alright. So that was my second question. We're not using the enterprise version that isolates your data from the general public.

Yeah. And that's and that's where we we need to get to is and part of that is this policy is Citi approved. And so once we have this, then we're gonna look at getting the enterprise version of it. So we want we want to we get we get requests to use different AI platforms all the time. You know?

And then every every vendor slap an AI on top of something that's machine learning. And so what we wanna do is instead of just, hey, go into chat GPT or this GPT or whatever, we wanna keep it in Copilot. And and and for that that reason there, if it's in Copilot, well, then all of our data loss protection policies

are in place.

Like you just mentioned, Paul, it isolates it from getting out to the general public. And so there's a lot of benefits to to having the paid version of Copilot. Right now, we just have the basic stuff that's tied in that all of our Microsoft products. So

we use Teams for our meetings and stuff like that?

Yes. Okay.

Yep. So then for the tools, like the auto record and summarizes meeting, do you disable those?

For teams? Yeah. We allow when you set your meeting, you can set it to auto automatically record, and then it pops up that it's recording your meeting. So from a from a Teams recording and transcription, we do that is something that we allow users to do. So, like, all these CAT RMS meetings, I set them to automatically record, and then I send a link out to all the team members. Like, hey. You can go back and rewatch this thing. You know? And I need those too. I need the transcript and the roll call. So

So is AI is the the AI that's built into Teams? Are you using that to auto generate summaries and transcribe the meeting, essentially?

No. I have not. It Teams will automatically transcribe it if you record it, but that's not like using Copilot. That's just

The paid version of Copilot, I think,

will do some of that.

Yeah. Yeah. Because we have it at work.

Yeah. Now I can open up Copilot. So here's a good example of what Copilot's doing now. I was working on an one of these one of these cover sheets a couple weeks back, and and I decided to just open up Copilot and just start seeing what I can do with it. And when I as soon as I opened it up, it said, what would you like to do today? And it had, you wanna work on the cover sheet for Winborn Consulting as an option. And I clicked on it, and I'll say a good 80% of that cover sheet was was generated by Copilot. Hey. Do you want to make, you would you like a suggested motion? Would you like to do this?

You know? And and it goes, you know? So but people are using AI out there now. You know? And it's just as you know, the danger of that is the danger of shadow IT, which we've been combating forever. Right? You don't you get a no from IT, then we'll go figure out how to get it done. And at least with how we're set up now, when someone goes to an AI site, it's automatically blocked. And it and then they have to click allow. And so when they click continue or allow, whatever that button is, then this is all through Defender.

Then it comes back to me, and I get a notice that they would like to use it. And then I can go, I can look at it in Defender for Cloud as all the different cloud providers in there and then ranks them on, you know, their security posture. Is this dependable, trustworthy? Is this a security risk? And so I can go and based on that, I can allow that one to go. And so I'm seeing that people are are going there. I'm getting emails. So we need to we need to get moving sooner than later. And so but I also wanna make sure we have a good policy.

What kind of things are people just out of curiosity, what what kind of things are people trying to use AI for? Well Give me some examples if you can.

A lot of people just replaced Wikipedia and Google. Yeah. Oh, yeah.

No. It's it's just a more efficient Google file. A lot of that. Let's see here. Let's see. Here we go. I'm working on a PowerPoint presentation, and my friend who owns a tech broker company in Wisconsin, said they are using this product called Claude. And then she this person that's requesting this sends me a link. Here's the big difference between Claude AI and ChatGPT. And, you know, so there's someone Okay. Probably trying to generate some images for PowerPoint and that sort of thing. Yeah. That was one that I remember right off the top of my head, so

I was

able to find

it quickly. Yeah. I was just care I I was just curious, like, if this was

Well, everyone's using chat to make sure their emails make sense. I'll say HR use chat GPT to tell us about the benefits fair that we just had because it had all the little icons instead of bullet points that, you know, that chat The tells. You. What's that?

That's one of the tells.

Oh, yeah. I'm like, alright. Well, Julia was using ChatGPT for this.

Yes. I can tell you it's being used illegal for lawyers who are receiving, say, a 150 page brief on something. They can pop that in there and just say, me a summary. Yeah.

Yep.

Boom. Brandon, that's excellent that you brought that up. I'm I'm very impressed with your your thoroughness on that.

Do you want me to tell

you more about that?

That's also really useful for any contracts. I mean, that sort of stuff. It just and finding information in instantly rather than spending hours combing through contracts, especially if you don't have an aptitude for that sort of thing. I have a few questions for you on this.

Yeah.

Is this intended to kind of restrict use really harshly until you get Copilot

functional?

Is that

behind this?

I think it's more to to get them to get peep before we get them trained. Copilot will be a huge benefit, but, you know, that's a fine line with, you know, with any policy is if you put it too strict, people will try and figure out how to work around it. So, really, we want the guardrails on this. So a lot of this language came from the city attorney too. So that's like the I would say the shells and shall nots.

That's that's Brian's language for sure. But yeah. It's it's fine and it's fine and a balance. I think every good policy has a good balance of like, yeah, we can use it, but these are the rules of using it, you know, and these are things you need to be aware of. You can't just copy and paste everything out there. You can't just, take its return and copy and paste that into an email and hit send. You gotta proofread it. There's even, you know, reference to, hey. If this isn't your work, you should cite it. You know?

And chat GPT, you can ask for citations if it's if it's quoting someone. And if because if you don't, how do you know? So Yeah. The reason I ask,

I mean, some of the language seems to be really restrictive. Like, for example, asking permission each time you wanna use it before you use it sort of thing. Mhmm. So it's it feels a little awkward. And if you read it by the way it's written, you know, it's I mean, then it's just those things I that seems to be a little bit unmanageable. Yeah. You know? And there's I think there's maybe two or three statements I saw that are kinda like that. One or the other one was encompassing all users. Does that include employees only, or is it everybody?

And then, you mean, like, contractors, everyone else? That one's worded a little awkwardly, I think. Yeah. And then the other one I saw was

let me think. I have some

notes here. Oh, I was curious about the HR use of it. If they don't use any PII or if they, like, redact it, I mean, it it kinda I think you say it's outright not usable by that department. I was curious if that is something that will be relaxed at some point. Because a lot of times it could be used to determine if AI created, you know, part of their resume. I mean, they could do research on Mhmm. Things. You know? And so that's why I think there's a lot of uses for it. But that's why I was real curious about some of those restrictions, and I asked my original question.

Yeah. Like I said, I some of this some of the restriction language is is gonna come from is coming from the city attorney. We need to find a good balance point. That's why this is this is why it's a discussion Sure. Item, not a not an action item tonight. And because I need feedback like you're giving me, Brandon. Like, this is good stuff. Like, if you're, if you're seeing something that doesn't quite make sense or if you think it's too restrictive or even, you know, too liberal. You know? Like, we need to find that good balance with it.

Well, I was thinking also, like, chat GPT, for example. I mean, most of the uses of chat GPT are for what you had said, things like email, correcting grammar, spelling, I mean, all that sort of stuff, maybe composing some small things. And again, to me, the way this is written is that you have to get permission to use that prior to using it every pretty much every time. Unless you say, okay. It's usable all the time because it's been approved for this, this, and this.

So it's just I'm just trying

to think of it from a user perspective. Yep.

No. That's good.

Yeah. And then my last comment is sorry. The last one is the definition of AI. You said you restricted it to generative, but there that's only one aspect of AI. I mean, you have analysis tools that are not really technically generative, maybe financial tools, I mean, other things like that. So that would be something else to think about. Yeah. That's defined there.

Alright. That's it. Okay.

Bob, do you have a ballpark cost of the professional version for it?

No. Okay. I don't. That's fine.

I know what the higher ed cost is. Yeah. So it's it was until December 1. It's 369 per user per year. Microsoft is changing for the higher ed version and knocking it down to $218 per user per year December 1.

Two eighteen.

But that's higher ed. I mean, that's the higher ed.

Is that, like, a permanent change to their license for higher ed?

No. It's hot off the press. We got a I found it today, and then our software won probably go. So I but that that's higher ed price.

So Yeah. Well, that's work. It was $30 a month per user.

Yeah. That's what Google is, Gemini.

And

that is there's no EDU discount on that for our k 12. And so basically, it's pretty expensive.

Yeah. It is. It is.

But it is I mean, having having your data isolated to your enterprises is a huge thing.

Mhmm. So Yeah. I think I think that's really the tip of the spear, so to speak. And, you know, I work in the industry to some degree. I'm not a not an IT security guy by by trade, but I was at a cyber security seminar earlier this year, and this this topic came up.

And I I guess maybe maybe I was being naive, but I never considered the fact that when you put stuff into an AI engine or whatever you wanna call it, that that information then became the property of whatever AI software you're you're using. And, again, I don't live in the security space as a primary job, but that never occurred to me. And even stuff like, you know, we you know, my company being in finance is pretty secure. We have Copilot, and Copilot is our our only acceptable AI platform by by rule, by decree. You know?

Even something like Grammarly, which I've taken for granted to use, you know, over the years, I never considered that when I throw a paper into Grammarly, that Grammarly, you know, retains that data of whatever I put in there potentially. Yeah. And so, you know, this is it's I I think from a from a typical user standpoint, you know, even an IT guy like me or someone who works in at least the IT related space never considered that. I'm sure a typical user who's not very tech savvy is not considering that at all. And I guess that's the the training part of it, but that was really eye opening to me, you know, because I never thought about it in that in that way or why this becomes an issue If you put, you know, stuff that shouldn't be public into that engine, it's no longer public.

You've shared it with whatever organization is running that AI engine.

Yeah. I mean, this is the these are the challenges we face with every emergence emerging technology we've ever faced and that's that's coming at us. Like, how do we secure it? And then how do we use it? And how do we use it right? You know, because you you think about, like I mean, the bad side of AI and what how it's gonna be used to, you know, exploit people and all that other stuff is, you know, the Prince of Nigeria's emails are gonna be a little bit easier to read now. Right? They are. You know, it doesn't matter what what country you're in. AI is gonna give you a properly written email.

So you have you have that stuff going on, but then you just see, you know, being able to, do all the deep fake video stuff. You know, I was just reading an article on how two two hackers, stole an identity from, that these hackers were from, I wanna say, maybe, like, Taiwan. And, they stole, the pictures and identities of a couple of gentlemen from Mexico, and they were doing job interviews. Oh. Serious?

Like, video job interviews. And and then they moved a certain way, and the mask sort of came off. And then I'll and then then didn't go back on, and that's how they got caught.

Yeah. Oh, jeez.

So Sure. So the problem you have there is, like and, I mean, we don't do this, but you there's companies out there that that's they they hire people remotely, and then they work remotely.

Right.

And if you can get into a company, you know, but with someone someone else's identity and you've done video interviews and all that stuff, that's nuts.

One important distinction I wanted to make with the data and is most of the engines, if you pay for it, and they have terms that tell you if you pay for it, it's private. They may use it. Or they may use it for training, but they're not gonna use it as, you know, any sort of profiling or any sort of retention. So that's something and most people don't know that. They just go and they look at the new free tool because they're coming out like gang busters Sure. Right. As usual. And they use it, and they start putting their stuff in. And, you know, that's that's kinda where the issue lies.

Yeah. Well, that's the that's why they give it to you for free Mhmm. To start using. And then after after you get used to it, then you so you were once doing for free to make their product better. Now you're paying them to make their product

better. Right?

And you are the product.

Yeah. Are we Let's be done.

I mean, is this this policy kind of putting words on paper to support what we're already doing to some degree? I mean, we've been aware of this. It sounds like you're already doing a certain level of management of this. And I guess, are we I mean, I guess, I'm kinda questioning, you know, what are we doing now? What does this policy add to what we're doing now?

And then, you know, is this laying the groundwork? I think you kind of already answered that, but is this laying the groundwork so that we can specifically say, these tools are approved. These tools are not. Don't use anything that's

not approved. Yeah. I mean, I guess, you know, it's called the AI acceptable use. Right? So you can use it, and here's what's acceptable, that you use an enterprise version of Copilot, that you don't put confidential information in there, that you do training, that you understand, what you're what you're doing, and you're not just, hey.

I think this is gonna make my job easier. So, like, Brandon reads it, and if he and he's looking at it with the eye the eyes of an end user, and they're like, well, this sucks. Yeah. That I that's what I don't want. Yeah. I don't want people trying you know, people are already trying to figure out how to use AI that's getting blocked by our systems. You know? So what I don't want them doing is that, and I don't want them trying to well, you know what? I'll just do this on my home computer. Mhmm. You know? And then I'll forward an email to myself.

Yep. Yep.

You know, I don't we don't want emails from work to personal going back and forth because they can't get to it here.

We want them to

be able to use it. We want, you know, we want people to do innovative stuff here.

Yeah.

Right? I always say just because we are the government doesn't mean we have to act like it. You know? We should be innovative. We should be, trying to figure out how to serve our constituents better.

And some of that is, you know, we have a lot of time invested in doing, tasks that might be easier done or might could be accomplished by AI and save fifteen minutes on the day to go get to that other task that, you know, of all the hundreds of tasks you have awaiting you every day. Everyone's wearing a ton of different hats around here. And if this if a tool can help them accomplish one so that they can focus on the other or just keep their head above water, whatever the case may be. You know? So acceptable use, I think, are the two keywords here is what we're looking for out of this policy, and this is the the first kick at it.

I think I mean, I read through it and I

okay. I

I read through it and I didn't have any, you know, significant objections, but I do understand what Brandon's point is, you know, especially, like, when we look at something like four four e, users shall report all AI use to their supervisors prior to use. Yeah. That's I think that's probably one of the bullet points that you were kinda calling out where that that seems a little draconian, you know. Yeah. I don't know what that means in terms of

you often say, I'm gonna Google this, just so you know. You know? And do I have to

do that every single time?

I'm topic change. I

guess I guess I guess the point of, you know, at some point, do we have a list of approved things that, you know, they can just use because we know they're approved as opposed to, you know.

Yeah. Well, that's why I I you know, you start with a copilot is the approved one. And then this is why it's approved, everyone. Right. And why we have to do this is so that when you're putting data into this system, it stays within our ecosystem. It's not available to the general public. I mean, I think that's create. Right?

That training point is Yeah. Is key, I think. Because, again, you know, that was eye opening to me. I didn't think about that. And I think, you know, again, I think the majority of users at the city probably don't think about it that way either.

Mhmm. Well, here's another thought. Was just it was just kicking around like, wow. All city communication business is public record because

it's Mhmm. Correct to some degree.

Well, with there's there's exceptions to it. But Mhmm.

If somebody does, like, an open records request, can they request the AI information?

Well, now it'd be the other thing too. Can we produce the record? If I don't know what what you're using, I don't know that that record exists. So because when open records requests are gonna come in, they're gonna come in to someone. They're gonna filter through the attorney's office, and then ultimately, this guy is gonna be doing some e discovery.

Yeah.

And, you know, and so across our three sixty five tenant, you know, we have we have a seven year litigation hold on everything. So you delete it. It's out of your view, but it's still discoverable. If you're throwing stuff in the chat, I can't I can't I don't know if that record exists, so I can't go get it. So there's that there's that part of it. It's really

But then I'm just thinking, wow. That could be

But then but then there's this the side of it, like, oh, well, we don't know that that record even exists. Well, if that record does exist somewhere and we're not producing it, well, then now we're and then they find out that that record existed and we didn't produce it. That's a whole another conversation there. One, it it looks really bad. Like, oh, well, they're hiding information. You know? And then two, well, they don't even have control of their data. Uh-oh. That's

a, you know, that's a really good. That's a But with or for thought.

Yeah. Mean, without

AI, you're still you're still in that pickle, you know.

What's that?

I say with or without AI, you're you know, someone comes in with a full request and you're still in the same predicament.

Yeah. I decided I'm gonna use Google Sheets instead of Microsoft Excel. Well, we're not a googlecom you know Google Sheets. Customer. Yeah. So we we can't we we have no governance over that. No. Right. So and we actually we ran into that issue just a couple years ago. Someone is using Google Sheets for something, and then, and then we found out about it.

And it's like, well, you know, you have people's personal phone numbers and addresses in here, and and it's not locked down. And so we fixed that because my address is not a public record. My phone number is not a public record. Oh my goodness. We've all signed up for something. And you continue to sign up for that.

Button for punishment.

But yeah. So there but to your point, you know, there the bulk of what we do is discoverable.

Yeah. Because that's I guess that's an interesting thing, and maybe that's a Brian question as far as what what retention of work product that's generated in AI needs to be Yeah. Well, that's just potentially, you know Yeah. Because, like, you know, I've had records requests and, you know, for me, it basically is email. Yeah.

That's my primary, you know, thing that I have any records that need to be, you know, disclosed because that's the majority of how alders communicate Yeah. Or direct phone calls. But that's that's, you know, that's not a there's not a record retention there. But if, you know, that's different for staff work where there's Yeah. Potentially work product.

We have a retention schedule. So, you know, a rule of thumb in in our world is everything is seven years. That's kinda rule of thumb. But there are some things we only have to keep for three, there but are some things we have to keep for twenty five. And there are some things that we have to keep for a lifetime. And so we we do have a retention schedule. And the people that need to keep those documents for 25 or a lifetime, they know that that's because when you come down to it, you're the custodian of of your records here. Every individual is. IT is not the custodian. We we we just give you the tools to help find that stuff when you need it.

Mhmm. And and we're not just giving you the tools. We're the ones finding it for you, but we're not your we're not your custodian. So and then there's other things too, like data backups. You know, if we don't have that backup system anymore, we don't have a way to retrieve those backups.

We don't have to keep those backups. So because we went we went through that, like, years ago. We got a letter from the historical State Historical Society. Like, no. If you do not have the system to to retrieve those records, you know, and it's a proprietary format like all backup systems are, then you don't have to keep them. Because we're switching backup systems. We're like, what are we gonna do with all these tapes? You don't have a tape drive? You don't have a system to get the

you know, what do we do with it?

Give it to the library.

They do. Well, you can't.

I can't.

Because what if a municipal court has a juvenile record sitting on that thing?

Well but then if that's the case, they have a case number. They don't have the juvenile name. That's what I've been told by the police department.

Yeah. But I mean, that's just a alright. So maybe that wasn't a good example, but it was a great example of like there could be sensitive information on there. And so you can't just say, well, here, but Right back over it. So, I mean, it's a yeah. It's interesting discussion. So, I mean, I really read over this thing and give me your feedback. I'll certainly be going back over this this conversation to get all of Brandon's feedback and start looking at it that way.

Yeah. Somebody out of coin for AI, their version AI, can of worms. Is that what it really boils down. The rabbit hole. Yeah. Yeah. But there's that's even better. Yeah. Yeah.

I think it's definitely something we need to get in place because, you know, it's not going away and You know? Folks are already aware of it and using it. So Yeah. We have a you know? If we have a direction that we're trying to, you know, put in place, whether that's copilot or, you know, copilot plus some other things, This lays the groundwork for that in my opinion.

Did did I read this right? Foxconn the old Foxconn down the road, Pleasant Prairie area. Microsoft Copilot is gonna own a big portion of that.

Microsoft is building a data center down there.

Yes. Down there. Yeah.

Yeah. Believe that was shut down recently. They did everything.

No. They wanna

Was it shut down and then they moved north?

Shut down. They moved north. Yeah. They moved north. Oh.

So they just got north.

That's that's the Microsoft one? Yeah.

I don't know. I just got an email from a recruiter that they're trying to hire 450 people to work in that data center. So Sign up. Yeah. No

windows. No. Nothing. Terrible commute. But Yeah. Oh, yeah. Maybe a hike.

Well, that. But

Did give I you a sec?

So they're building they're building something somewhere in the Pretty

sure it's Port Washington.

I believe it.

Yeah. Just saw it on the news. Like, it was a big town hall meeting about it, you know, and We Energies was there, you know,

and Yep.

There's environmental groups that are opposing it.

I thought that was I could be wrong. I I know Oracle was looking at building 1 Yeah. Somewhere.

This is a this is I a I heard that too. This is a Microsoft data center. 700 acres.

How many? 700. Yeah. Over a square mile.

That's Wow. That's big.

One of my engineers talked to a Dell technician, I don't six months ago, and they were prepping to go with down to Pleasant Prairie for that data center. And he was talking about there's gonna be, like, 1,300,000 servers deployed in this thing. It would take a year and a

half to

deploy with a full time.

Well, then I think they said something this I could be wrong on this, but on that 700 acres, six buildings. Yeah. Those are massive buildings.

Yeah. Yep. That's huge. One square mile. Wow. Alright.

Little bit of green space in there, but

Yeah. Those two spaces, WeEnergy is saying that those two spaces are gonna consume more electricity than the rest of the state residential unit.

No way.

Yeah. That's what they're saying. You know, so we better start building more windmills.

Yeah. Well, that's every community along the Great Lakes is gonna be going through this. Yeah. Because Because they

need the water. Oh. Cool. Sure.

Yeah. Even if they're building their own nuclear power plants at

the site, they still need water. Yep.

Yeah. They need even more

if they're building towers. Yep.

Okay. Chris, do you have the feedback you're looking for?

Bring it. Arrange this in. We can talk after the meeting. It is a

it's it's a revenue. Yeah.

Oh, yeah. Any other comments, questions about AI? So I think we'll we'll be seeing this back before us in the near future, I assume. For approval? Mhmm. Yes. Or for more discussion. Yeah. Hopefully, approval at some point.

It'll be back in December. I won't be here. Greg will be here in December. And then we'll we'll we'll have a contract to, for something chief was alluding to earlier.

Okay. Okay.

So I I So

we are planning the December meeting. That was gonna be a

question. Yeah. Okay.

Alright. You'll pay for Okay. Which time do you use it? That takes us to item six, communications and referrals. Anything else? Nothing. So plan a meeting in December. Else from the board? Alright. Item seven is adjournment. Any objection to adjourning? Alright. Goodbye. Hearing and seeing no objection, we are adjourned by unanimous consent. Seven thirteen.