Community IT Innovators Nonprofit Technology Topics

Nonprofit AI Governance Tips with Nura Aboki pt 2

• Community IT Innovators • Season 7 • Episode 43

Share episode

Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.

0:00 | 27:22

In Part 2 of the Nonprofit AI Governance Tips webinar, Carolyn Woodard and Nuradeen Aboki, Senior Consultant who has been helping nonprofits explore and adopt AI tools, cover the practical steps your organization can take to move from AI experimentation to intentional governance.

The conversation picks up with a closer look at what belongs in a nonprofit AI policy, who needs to be at the decision-making table, and how to make governance stick in day-to-day operations rather than just on paper. Nura shares a case study of a content-heavy nonprofit that built AI guardrails around their editorial process and came out ahead, and the two close with a question-and-answer session covering metrics, ethics, and the environmental impact of AI.

Haven't listened to Part 1 yet? Find it in your podcast feed.

This episode covers:

  • A good AI policy addresses acceptable use, data handling, compliance requirements, vendor vetting, human review, and staff training expectations — and it needs to evolve as the tools and your organization do.
  • Cross-functional governance works best when leadership, board oversight, IT, legal, HR, and end users all have a seat at the table. Research drawing on a national survey of 180 nonprofits found that organizations where staff and board co-developed AI principles launched 12 times more pilots and scaled AI more effectively.
  • One nonprofit built AI governance around their editorial workflow: updated style guides, required human review of every AI-assisted draft, and targeted prompting training. The result was faster writing without sacrificing voice or accuracy.
  • When things go wrong, the first step is a calm assessment: figure out who is using what, what went wrong, and whether the root cause was a training gap, policy vacuum, or misconfigured setting.
  • Making governance real means publishing your policy internally, raising AI use in staff meetings regularly, and creating spaces where people at every comfort level can ask questions and share concerns without judgment.

Resources Mentioned:

_______________________________
Start a conversation :)

Thanks for listening. 


Community IT Intro

Thank you for joining this Community IT podcast, part two. You can find part one in your podcast feed if you subscribe wherever you listen to podcasts.

Carolyn Woodard

Welcome to the Community IT Innovators webinar on nonprofit AI governance tips. So my name is Carolyn Woodard. I'm the outreach director for Community IT. I'll be the moderator today.

Nura Aboki

Hello, everyone. I'm Nura Aboki. I'm excited to be here. As Carolyn mentioned, I've been helping nonprofits explore and adopt AI tools over the past year or so.

Carolyn Woodard

This one is also just a really quick uh sort of guideline of what a good AI policy should cover. I'm gonna include this link in the chat of our the acceptable use uh template that we have on our website. That's a free download. You can use it as a starting point. Um, but

Carolyn Woodard

The things that a good AI policy should cover is that acceptable use, um, the data handling rules. If you have compliance that you have to be aware of, like HIPAA or you work with you know sensitive communities or children, like what data you're allowed to just through your regular IT policy that you're allowed to use, uh, human review requirements, um,

Carolyn Woodard

The vendor vetting, like this could be its own webinar because it is so opaque and so difficult to find the real information about what the different AI tool vendors, the you know, big six tech companies are doing, um, who is approving the new tools, uh, training expectations. So, what you are expecting in terms of your staff, what training you're going to provide, what training you want them to have as they're using the new tools. Um,

Carolyn Woodard

So all of that is something that can go in your policy. I would say your policy definitely needs to be a it needs to be something that is evolving as you're evolving. You're it's changing so fast that you need to be able to be flexible with that. Um and then

Carolyn Woodard

Nura, I think we're going to talk a little bit about who is at the decision-making table and what formats can kind of work for that.

Nura Aboki

Yeah, uh let's talk about that. Uh, you know, it uh we strongly recommend a cross-functional uh team or a committee that brings together uh various perspectives. So

Nura Aboki

You'll want leadership, certainly, uh, because leadership supports uh is essential for resource allocation and culture. Uh then you want to also look at possibly a board member or board oversight, uh especially on risk and ethics side of things. You know, this has been mentioned in the chat quite effectively. So a strong board or strong board member, that board oversight will help with that, addressing the ethics uh side of things. Uh, getting your board on board uh to speak uh is uh helpful uh for setting the tone and maintaining accountability.

Nura Aboki

Then your IT team or provider uh definitely needs to be at the seat uh at the table. Uh they understand the systems and security implications. So you have that advice, uh, IT advice in terms of implementation.

Nura Aboki

Legal HR and on compliance roles uh are also important there. So into when it comes to governance, since governance is all about policy and people, HR can ensure it's integrated into training, into hiring, and legal can lack compliance issues if you have some.

Nura Aboki

And more importantly, representation from teams uh that will use AI, basically, everyone that's going to use AI. So the end user should have input uh to share concerns and uh practical needs. Uh uh more the more diverse uh the committee is, uh, the better. Uh you'll catch more potential issues and uh gain more buy-in. In fact, research uh by uh TechSoup and Board Source, uh for instance, uh found organizations where uh staff on board uh co-developed AI principles uh launched 12 times more pilots and scaled AI more effectively. Yeah, sorry, go ahead. Yeah, go ahead. So I was gonna just uh wrap it up here. Having the right people involved not only uh produces a smarter policy, uh, it also builds trust.

Carolyn Woodard

Yeah, so um I shared the link in the chat as well from uh board.dev, which is this great organization that helps with developing nonprofit boards around technology. And um, yeah, they it's just this really wonderful um report that they did. Uh so I'm gonna keep moving us along. Um, you know,

Carolyn Woodard

You were gonna share another case study with us about kind of how individual the policy might need to be.

Nura Aboki

Yes, uh this case uh study case highlights a nonprofit uh that did a lot of things right uh by starting with governance in mind. Uh

Nura Aboki

This organization produces plenty of uh public-facing content. So think of reports, articles, newsletters. So they wanted to leverage AI uh for initial writing and editing assistance. Uh right away, they realized this was a high-risk uh area because anything inaccurate or off-brand might reach the public and harm their reputation.

Nura Aboki

So they set up a governance approach uh around writing using AI. Uh they updated their writing style guides and uh standard operating procedures uh to incorporate AI. For instance, they gave staff guidelines on how to use AI to draft content uh that will sound or that still sounded like the organization's authentic voice. Uh they required every AI assisted draft uh goes through normal editorial review process. So you can see human in the loop, ensuring that a person checks facts and tone. Uh,

Nura Aboki

They even did uh targeted uh training sessions on how to write with AI, uh meaning how to prompt the tool with uh their specific style and how to edit its output uh to meet their standards.

Nura Aboki

As a result, they were able to speed up some writing tasks without losing trust and consistency with their audience, uh, their audience expects. Uh so the key lesson here is effective governance is uh specific to your nonprofit organization's needs and values. You know,

Nura Aboki

That's very important here to understand your needs and values. Uh, in their case, content uh quality and voice uh were paramount. So they made sure that their AI policy and training honed in on that. And because they started early with the guide with this guide rails, uh their staff felt comfortable and confident using the tools.

Carolyn Woodard

Such a great example of how specialized your your policy might need to be and kind of areas where you might be able to make that you know start, get started because you know that they're higher risk or very important to your organization. Um so now

Carolyn Woodard

I wanted to give you a prompt for the chat. Um, has your nonprofit had an AI-related problem or close call? You can drop a yes or no in the chat and um give some examples. Um,

Carolyn Woodard

Some examples we've seen are like a permissions sprawl. So you have um, or I guess permissions that are, you know, maybe not very locked down, and suddenly you have AI and it's giving people access to documents that they have permissions to that maybe you didn't know they had permission to it. Um something due to implementation that wasn't guided or intentional. Uh,

Carolyn Woodard

As Nura was saying, the the issue of, I mean, they didn't have the close call, but they knew in that organization that they were gonna want to make sure that their AI-assisted publications, the editorial process was just as rigorous as it would be.

Carolyn Woodard

So um we're seeing some, I want to see in the chat a couple of things. Um, most people are saying mercifully no, not yet. Um generating AI slot pictures or misspelled words and slides. Yep, that happens. Um, I had one that I had put on our website and I still had the little prompt and I didn't realize it until the next day. I was like, oops, I put that in there. Um, although I'm very transparent at the end of my AI-assisted um blog posts, I say at the bottom I used AI to um to help me with writing. So um no one's showing up for meetings anymore because they all send their AI summary bots. I love it. All right.

Carolyn Woodard

I'm sorry we're gonna have to get, as I said, we're gonna go pretty quick through some of these last slides. Um, but Nura, we wanted to talk about where you are in the process and maybe some next steps.

Nura Aboki

Yeah, uh, so we know folks uh attending today are in different places with AI. So we want to provide next steps uh for three scenarios here. Uh,

Nura Aboki

If you're just starting, uh maybe you're a skeptic or have no clue where to begin. Well, here's the first thing you want to do is audit what tools uh staff are already using. So sometimes there's the cross discovery tools that could be added onto your computer system, so you can ask staff to tell you which tools they are using. Uh, you might be surprised uh how they are using AI already uh under the radar. So a quick survey would help, or informal check will help you with that.

Nura Aboki

Then you start with the leadership conversation. Uh bring AI up to your leadership team and board or board if it if the board is accessible and then share some of today's insights to get uh support and assign ownership.

Nura Aboki

Then you want to leverage uh templates and resources. Uh for instance, you can download our AI Access Will Use Policy template, which uh was provided there. And also you can adapt your to your needs. So downloading is one, but really looking at what your organization needs and adapting it is what where the work uh needs to happen internally. And having uh even a simple draft could help you uh with those discussions and actions.

Nura Aboki

Then uh another, if you are building AI, uh then maybe you have already started using AI or you're actively piloting it.

Nura Aboki

You want to identify your governance leads, uh, kind of form that cross-functional team that we mentioned earlier, a designated uh owner, AI owner to manage the this initiative. Uh,

Nura Aboki

Then you want to draft and adapt that, uh, adapt the uh AI acceptable use policy or governance policy that will be putting together in writing uh now that areas that are covered within the acceptable use policy in terms of uh data or any oversight that is happening, and then refining it as you learn more about the pilot. Um

Nura Aboki

You want to build in training. Start with staff training uh as you uh deploy the AI tool and then plan to review your progress uh regularly. This uh has to have some frequency to it. It may be more than instead of having a single-year review, you may want to create time to do it quarterly or even monthly, depending on uh as you are going into implementation because continuous improvement is key when it comes to these policies. Uh

Nura Aboki

So if you need uh reset, then uh you would say, you know, you've you've already had an AI mishap or realize things are chaotic, don't panic, uh do an assessment. That's the first step. Figure out what's happening, uh, who is using what uh AI tools, uh, what went wrong and how uh bad it was. And it's important to understand the root cause calmly.

Nura Aboki

Then you want to identify what happened and why, and uh then that will help you kind of was you ask those questions. Was it a training gap, a policy vacuum, uh misconfigured setting? Uh getting to understand why will help you fix the issue.

Nura Aboki

Then you can reset your governance there uh based on what you find, tighten and create uh your policies and policies accordingly and communicate changes clearly to your team. It might involve pausing some AI use uh until guardrails are in place, uh, then relaunching them with better oversight.

Nura Aboki

The overall message here is wherever you are, take the next step. Uh even if even if it's uh small steps uh like drafting a one-page uh policy or asking your staff what they are currently using can significantly reduce risk and set you on the right path.

Carolyn Woodard

Yeah, this uh slide is pretty packed. Um, but again, I hope that it's helpful to you, and this will all be on our website as well at community it.com. But um, you know, kind of

Carolyn Woodard

Ideas for making it stick that uh we've talked a lot today that governance isn't a one and done. Your IT policies need regular review and updating all of them, but especially your AI policy. Um,

Carolyn Woodard

Staff training is definitely an ongoing issue for these types of tools. So update your onboarding, um, but also build in learning time and a cadence and style that fits your organization and have regular reviews and check-ins with your staff or with everyone, all staff, to ensure the training is working for your people. So people have so such different learning styles. Some people like to go off by themselves and just learn their tool, just play with it. Other people really need something more structured. They want to do it together.

Carolyn Woodard

I think one thing I've been hearing a lot is that the more we learn AI together and become AI literate together, the more unified our voices in the philanthropy sector can be on what we want these tools to be like and how we want them to work and what ethics and safeguards we want and put in place. Um, and also uh

Carolyn Woodard

Don't forget to build in some checkpoints. Um, so pilot a lot, build multiple go no-go points into your decision making for big AI decisions. And remember that right now your AI tools are being implemented through a lot of little decisions as well. So just kind of wrapping your head around the whole the whole issue that is uh AI at nonprofits.

Carolyn Woodard

I'm gonna share some additional resources with everyone in the uh chat, and it will also be in transcript and on the website. Uh, some of them are from our website already. We have an AI library, we have a governance library, and we have the nonprofit AI podcast on Tuesdays, which is where we talk a lot about some more of the ethical issues and environmental issues that are coming up. There's some good frameworks, there's that um uh literacy link to the Department of Labor article. Um, so I hope that these are helpful for you. And um we're gonna move on.

Carolyn Woodard

I think we do have just a minute or so for a question. And um, we had a really good one at a registration. So I hope that it's okay if I share that.

Carolyn Woodard

I'm gonna ask you, Nura, uh, how do we ensure AI governance frameworks just don't exist on paper, but are actually implemented in day-to-day operations? And I feel like since you've been working with organizations, you've probably seen the good, the bad, and the ugly around this. So if you could share with us a couple of tips.

Nura Aboki

Yeah, uh that's the what are the key challenges. Uh but

Nura Aboki

I think uh what I've seen with success is leadership recognizing how important AI is to their mission and ensuring that stakeholders across all departments and teams have a continuous kind of review and communication with their staff around the AI use in their organization.

Nura Aboki

So whether it's having the governance policy published in their internet and periodically in their staff meetings just bringing up progress and the use of AI, that has been effective. I guess it's new and people are interested. And some uh use cases where collaboration has happened, and a staff meeting is actually a good opportunity to share. Hey, this is what we're doing with AI, so that others that are maybe skeptical or have not yet give will be given the opportunity to just be aware that the progress is being made and there's governance around AI. So leadership buy-in and making sure that in meetings or in the communication is happening regularly, more than it used to be in the past.

Carolyn Woodard

I heard it also said that um it's important to have spaces where people can talk about it and have that be, you know, welcoming and uh create a sense that we're all in this on the team together, right? Your nonprofit, and people are in way different places. Um, people have really strong opinions on AI for sure, on their tool that they're using that they love and uh what they're using the AI for, whether they should be using it at all. Um, and that

Carolyn Woodard

It's important to have those kind of learning spaces where everybody is welcomed at whatever point they're at, so that you can really talk about it transparently. And that saves you from finding out later on down the road that somebody just has not been on board with this whole thing the whole time, or that somebody has gone off in their own direction and maybe purchased and spent a lot of money on something that, you know, maybe isn't secure for your organization or doesn't align with your values.

Carolyn Woodard

So making that kind of space where people can be open and have questions and appreciate each other's points of view, I think is also something that's um particularly important around AI.

Nura Aboki

Absolutely. That's an excellent point there. You know, that regular space uh that is something regular, people know they can jump into that space and have those candid conversations without being judged, you know, about their position and whether they are for or against AI, is important.

Carolyn Woodard

Yeah, yeah. We got one more question in in the chat, which is uh what metrics have you found helpful when organizations are developing and measuring their AI governance? That's a good question. Yeah, that's so early in the game.

Nura Aboki

Yes, you know, we're in the mostly, I would say, an experimental phase uh at the moment. Uh, but organizations actually can look at uh key metrics when it comes to whether, you know, is it really truly saving staff time or staff just still experimenting uh with these tools? So uh

Nura Aboki

If you have defined goals, uh that is one way you could actually track governance, uh, kind of how governance is working. Um, one area that I've seen uh is through the rollout process, that definition is uh presented in terms of the goal for using the AI. So we ask the question of why do we need to use AI and what are the goals we want to achieve using AI that would help you then develop metrics that you could use to measure the performance of the user of AI, knowing your goals. Um,

Nura Aboki

Some goals uh could be automation of workflow uh processes uh between teams. Uh, some could be an example where uh this organization had to develop a process for the nonprofit to be able to write in a human-like using AI, human-like uh process using AI.

Nura Aboki

So goals uh really help you define what your matrix uh would be uh ultimately.

Carolyn Woodard

We have another question, which is how do you take into account the ethical and environmental impacts of AI in your recommendations? And I know you've had organizations who have been asking you about that. So do you have a quick answer? That's like a whole webinar.

Nura Aboki

Yeah, it's just a whole yeah, yes, it is a whole webinar.

Nura Aboki

But indeed, you know, uh core values are very important, and we cannot dismiss uh that from people. A lot of people are working in a nonprofit organization, they believe in the mission, they believe in the cause. So having an understanding of the environment itself, what is environmental impact? Some of them are quite exaggerated, some of them are actually real here. So

Nura Aboki

I think that education piece, that literacy about the true impacts is needed in order to make a true judgment call. But you can see uh from certain examples that we've seen that the true environmental course is kind of hidden and it's not quite transparent. So having to get that information is really critical uh to helping you decide whether the use of AI is actually worth and it's not against your permission and your ethics that you totally understand.

Carolyn Woodard

Do you mind elaborating on the exaggerated impact?

Nura Aboki

Yeah, so uh so for instance, uh if if you think about energy, you know, how what energy is used? Ideally uh the way it's communicated to us is when I write a prompt, I am using a lot X amount of energy. My prompt is X amount of energy. But if we look at it holistically, this prompt is not just one person writing this prompt, it's not just my core responsibility, but as a whole, the entire world that is using AI, we want to understand how much of that is truly impacting the world.

Nura Aboki

So it could they could say if you like, write a long, long prompt is actually consuming X amount of compute power. But as a whole, how much compute power are we actually truly using? So that is where having to understand those nuances would help you in defining whether this goes against your ethics. Uh there, because there are key nuances within when it comes to prompt engineering, for instance.

Carolyn Woodard

Yeah, so there's different, I mean, it's all it's complicated, right? What you're taking into effect. Are you close to a coal-fired plant, which is where your electricity is coming from? Are you using more uh renewable energies? Like where in the country are you? Where in the world are you? There's lots of different um complicating impacts. Uh, I will,

Carolyn Woodard

I can't find it right now, but I will put it in the transcript. There was a really good ebook from MIT that just came out a couple of weeks ago there where they did a lot of estimating of what the total cost of energy, and um there's some other organizations that are working on the water use issues. So I appreciate the question. I will share that um link when I can. I'll put it in the transcript. I was just looking for it now and I couldn't find it quickly. Um,

Carolyn Woodard

I wanted to go back over the learning objectives for today. So, what do you need to know about AI implementation? What is AI literacy? What is AI governance? What are good models to follow for decision making? What are some AI risks, and what are some emerging best practices from case studies?

Carolyn Woodard

Nura, I want to thank you because I think we covered all of this, even though we had to rush right through some of it, um, but we did touch on all of those things, and I hope we gave the audience something more to think about. Um,

Carolyn Woodard

I wanted to invite everyone back next month. We are doing another webinar, uh, which is uh, well, AI will probably come into it, but it's not about AI, because I, if you're like me, everything is AI all the time.

Carolyn Woodard

So uh we're gonna talk about Google Workspace, which is something that a lot of nonprofits are using and uh administer themselves, especially if you're a smaller nonprofit, your startup, um, and you just set up Google Workspace yourself. I set it up for my own business uh several years ago. Um, and if you did that, you are in good company.

Carolyn Woodard

We are seeing up to 50% of our new clients come to us with Google Workspace as they grow and they realize they need a bit more help to manage it. Uh, we are one of the only MSPs that has a lot of experience with Google Workspace.

Carolyn Woodard

And we are having our expert in that, Steve Longenecker, who is our director of IT Consulting, with us in June to talk about the differences with how Microsoft sets up security, how Google Workspace sets up security. Um, there are some things that you can do. Google Workspace can be very secure. You can do some things as you grow that will keep you secure now and prepare you to stay secure later if you grow to larger staff members, if you want to stick with Google Workspace, or if you're using it in a hybrid with some Microsoft as well.

Carolyn Woodard

So that is gonna be at 3 p.m. Eastern, noon Pacific on Wednesday, June 17th. And I will share that link with you now. You can register if you are on Google Workspace and are thinking about security. This can help you. And you can join us on Reddit slash R slash nonprofit IT management for more questions and answers with Nora for about the next 30 minutes or check in later, uh, check in anytime, and we'll be checking back on that Q and A thread over there to see if we can answer some more of your questions.

Carolyn Woodard

Nura, I want to thank you so much for joining us today and helping us with all of this expertise and all the experiences that you've had with our different clients, large clients, smaller clients, and uh different things that they're trying to do with AI and how they're managing their governance. So thank you so much for joining us.

Nura Aboki

Thank you for having me. It's a pleasure.

Carolyn Woodard

And to everyone who joined us in the webinar, I really appreciate your time today. You spend an hour with us, which is an hour you could have been doing something else. So we appreciate it. Hope this was helpful. Uh, it's a huge big topic.

Carolyn Woodard

We, as I said, we have lots of other AI resources on our website, community it.com. So I hope you can join us there or over on Reddit. And uh we'll be back next month with another webinar. Uh you can find all of our past webinars are in videos on our website as well.

Carolyn Woodard

So um just thank you so much for joining us today and um take care.

Community IT Intro

Thank you for joining this Community IT podcast, part two. You can find part one in your podcast feed if you subscribe wherever you listen to podcasts.