Transcript:
IAEM24 – Democratizing AI for the Whole Community – Suzanne Frew
Introduction and Purpose
Welcome. Thanks. I appreciate you guys all being here.
[00:00:04] Building Cultural Competency in AI
I’m going to explore this area today of, “How do we really build cultural competency in artificial intelligence?” If we’re looking at the systems now, what can we do to really build it for whole communities?
So, my question is, can we? And if we can, how do we do that where we can look at human-centric AI? And we’re looking at that because our mandate is to look at whole communities. And so, what I’m going to do is pose looking at the entire area of emergency management, and then how we can look at our AI models, and how we can build that in there, so we can really incorporate cultural wisdom.
Now, I’m not hearing this talked about very much. As a matter of fact, I’m not hearing it talked about hardly at all. But there is some space where I think we can build this in. If we’re looking at long-term recovery, if we’re looking at preparedness, and all of those places, where can we really build AI to work for us, to serve us, and to make those systems?
[00:01:05] Current Use of AI in Emergency Management
So, right now, we know AI is being used in emergency management. There’s the conference that we’ve been at so far, and there’s been one session after the other talking about AI, who is using it, and all the models that are being built. So we have a lot of different examples that are out there.
It’s being used in operation centers. It’s being used in preparedness. It’s being used in evaluation after a disaster—all of these areas. So there’s a lot of richness right now, especially where we are using AI. So that is the thing that I don’t think anybody’s questioning. We’re in the wild, wild west right now.
So, we are really trying to figure it out, you know, maybe 10 years ago, we were looking at social media this way. Now, we’re looking at AI this way. So, this is the thing that we know it’s here. This particular example was in Türkiye, where they were using AI in a search and rescue mission under a collapsed building structure.
So, we have a lot of systems in place right now. It’s actively being looked at. It’s actively being used and explored. For me, I’m coming at this with a beginner’s mind. I am surrounded by a lot of people who are doing a tremendous amount of work in AI—a lot of companies, a lot of individuals. And what I want to really push it for is saying, well, where is the cultural competency in that and how we can do that?
What can we do? So, I’m looking at what’s new. So, what are some of the collaborations and initiatives that are going on right now? Well, here’s one example that I think is pretty exciting, and I just found out about it a couple weeks ago. And I want to explore all of this area I’m really trying to explore.
So this one is the Global Initiative on Resilience to Natural Hazards through AI Solutions. This is something that came around about four years ago. There was a, there was a group that got together at the international level. They started looking at where we could apply AI to emergency management solutions, and they just met for the very first time in Spain earlier this month, and I’m really looking forward to hearing what’s coming out of that.
So, we know that we’ve got a lot of initiatives, a lot of collaborations that are going on. I’m posing; why can’t we actually do that even further? and looking at the cultural competency and moving it in the direction that we would like to do. we know we want to look at, the data collection, we want to look at a lot of these areas.
If we’re going to be doing it in support of servicing whole communities, then this is one collaboration that we might need to move forward.
[00:03:48] Challenges and Limitations
So, Challenges and limitations. Wow. That’s a big surprise there. We have all kinds of challenges and limitations. So, one of the biggest areas that I see of huge concern is, historically, the data that’s been collected has been very limited.
We are not gathering data, perhaps in the best ways that we need to be. We’re not reaching out to a lot of those populations that are disproportionately impacted by disasters in this country, in the United States, and abroad. There’s a lot of places where we’ve got a lot of real severe, climate change that’s taking place.
We’ve got a lot of areas where we’re not being able to gather a lot of meteorological data. There are small island nations where we are seeing this, but we don’t have the data collection that is going on. And with AI, if we can’t put it in there, we can’t get the solutions back. So we’ve got some really severe limitations right now.
And I think one of the things that we’ve got to figure out is, how do we go about overcoming what these challenges are? because we can’t move it further if we’ve got skewed data. Now I know that there are a lot of initiatives taking place, and I think we need to look and bring, some of those initiatives together.
[00:05:09] Importance of Language and Communication
So one of my areas of real strong interest is language and communications. That is something that, for example, is one of the areas in AI that I think is very promising. It’s also a big challenge right now. So, if we take a look at what AI is being used for in the whole language area, we’re trying to make sure we can go out in communications to get these threshold languages.
Because when we are trying to move out, alerts, if we’re trying to get information out, we need to get these different language sets out there built into the AI systems. What are we missing, though? We’re missing a tremendous amount of language that’s out there. This is particularly near and dear to my heart.
This is a photograph that I took in Belfast, Ireland, off of the Peace Wall, and they’ve had a lot of historical challenges there in looking at the Irish language. And in December of 2022, they finally got it put into law and recognized as a language. And there’s recently been a film that has come out called Kneecap.
I don’t know, has anybody heard of it, seen it? I think it won the audience award at Sundance. It’s a very, very powerful film about the fight for a language. And at the end of this film, it talks about the fact that there is an indigenous language that is lost, it dies every 40 days. I mean, that is a very powerful thing, if you start thinking about that.
Well, when we start looking at emergency management, and we start looking at what we need to do, we start looking at these AI systems, we need to get community collected data that really represents that community. And get that built into the system. Well, how are we doing that? That’s one thing that we can really look at.
We can figure out how we could build that into some of our programs moving forward, some of our funding cycles. So the language piece is very important for not only languages because a lot of it, we’re looking at these very standard languages and this nation is changing dramatically. The demographics are changing dramatically and we have a lot of challenges with the language.
So, we think about how do we get all of these other languages that are not your traditional norm languages. You’ve got a lot of dialects; you’ve got a lot of areas like this that we can actually build into our AI systems. So that is one area that I think is really important. Again, we’ve got to build a system that is going to be reflective of this whole community or we’re not going to be able to serve that community.
[00:08:01] Case Studies and Examples
So, I now want to talk about a couple of case studies. That I think is very interesting, and in the emergency management community, I think that that is something we should really look at. So one of them is something that has been developed out of the private sector, and it is by, everybody know Dove Soap?
the Dove Corporation has had a focus for a long time on “What is beauty?” So, one person’s description of beauty can be very different from one to the next. So, Dove has come up with an amazing playbook on “What is beauty in AI?”
And they have a new campaign called “The Code,” and in the code they have actually produced a playbook that talks about how we start building equity and inclusion in AI. And the playbook is chapter after chapter. It’s about inclusion. They have a glossary, they’ve got prompts, all of the ways that you can build that into the campaign.
That is a great example where they have absolutely directly addressed. This challenge in AI, and they’ve come up as a corporation with a playbook to counter that discrepancy. So that I think is a really important thing that we could look at is when I went through that playbook, I was pulling things out going, this is emergency management language.
Why can’t we take that as an example and move that over? So, their real beauty prompt playbook. I’m thinking about, hmm, where could this come over to what we’re doing? The next one that I think is extremely exciting is in Hawaii. There has been some tremendous work that’s being done. The University of Hawaii has been very involved with it.
It’s a multi-million dollar funded project and they are looking at traditional wisdom out of the indigenous culture in Hawaii and they are figuring out how to integrate it into health care in their health care AI systems. So, this is something that’s being done by multiple departments at the University of Hawaii, along with indigenous practitioners within the state.
And what they’ve done is they’ve looked at some of the cultural traditions and the cultural values in the state. For example, mālama, you know, caring for something. mālama, kuleana, which is the responsibility. And what they’ve done is they are figuring out ways to integrate this traditional care, the traditional aloha, the sense of aloha that is traditional to a lot of Indigenous communities and a lot of communities that might be of a different culture within our own town, our own, our own city, our own, you know, agricultural area, finding ways of looking at the ways that they’re doing their integration process.
So, I started thinking about that, and they, they have gotten tremendous amount of support from an international source, funding source. They’re actually doing that embedding and apparently very successful. I just found out about it recently; I’m coming at this with a beginner’s mind. I think this has tremendous opportunity.
[00:11:34] Proposing a Cultural Competency Playbook
So I started thinking, okay, why can’t we in emergency management put these two concepts, these two case studies, together and create something that is a playbook? I mean, some of the companies out there, for example, Deloitte, has done an entire, they’ve done a playbook right now for ethics and equality in AI.
But looking at some of the initiatives that have been going on and taking a look at some of the work that has been done, the initiatives and the collaboration that’s been done and create our own playbook for integrating cultural competency, which would reflect cultural wisdom and traditional practices, whether it is for a first nations community, whether it is for a specific community that is disproportionately impacted in a disaster and create playbooks in our communities for addressing whole communities.
And so that’s what I’m really proposing is that we look at how do we do that? How can we come together? And I would be very interested in collaborating with someone, an organization or whoever. I think there’s tremendous opportunity here to really move this. And build a whole community AI playbook for emergency management.
And that way we can build a foundation. We can look at the traditional wisdom that we’re forgetting a lot of the time. And a lot of times, we don’t have opportunities because of funding crunches and things like that. But let’s build something into a playbook.
[00:13:18] Conclusion and Call to Action
So, that is what I wanted to present these ideas today because I think there are some real possibilities.
And I welcome any feedback. I think we’ve got a lot of opportunity. We’re in the wild wild west with AI, and let’s just keep going, and we can build that in to building more equitable disaster prevention, response, recovery, all the way through that really addresses the whole community. So, thank you.
And if anybody wants to be in touch, I would love to hear from you. So, thank you.
