Can AI Change How We Deal with Disasters?

Deanna Lee

You're listening to Policy Minded, a podcast by RAND. I'm Deanna Lee. Disasters are becoming more frequent, more severe, and more costly. And that trend is intensifying the difficult challenges already facing emergency managers across the United States. So how could AI tools help them save lives, protect property, and reduce harm to communities?

A new RAND report out this week explores that question—mapping the landscape of AI-based tools that could help emergency management agencies and the organizations that support them. Today, we're joined by two of the study's authors, senior policy researcher Jessica Jensen and senior adjunct researcher Glen Woodbury, to discuss which kinds of AI systems may be most helpful now, the challenges of implementing those systems, and how AI can be used to create lasting improvements to disaster preparedness, response, and recovery.

Jessica, welcome.

Jessica Jensen

Thank you. Happy to be here.

Deanna Lee

And Glen, thanks for joining us.

Glen Woodbury

Thanks, Deanna. My pleasure.

Deanna Lee

Before we get to the report, I want to mention at the top of the show that this study was commissioned as part of the Markle Foundation's AI for Disasters and Emergencies initiative, or the AIDE initiative. Can you tell us what that is?

Glen Woodbury

Sure, the AIDE Initiative is an AI-for-public-good effort focused on helping emergency managers, government leaders, and communities across the country better prepare for, respond to, and recover from disasters. Markle brought together emergency managers and others from across the country not just to identify promising uses, but to help organizations adopt them responsibly. RAND's role, at least initially, was to map the tools that are already available, where they fit into emergency management work, and what it would take for agencies to use them effectively.

Deanna Lee

Okay, great. And we're going to talk about that work you did. We will link to the AIDE initiative in the show notes for people who want to learn more. So on to the report. What question were you trying to answer with this study?

Jessica Jensen

We actually had three questions that we wanted to answer. The first being something that Glen already alluded to, which is what are the actual AI-enabled products that are out there today that could help emergency management right now or in the very near future with moderate tweaking of those products? But it wasn't just mapping that landscape that interested us. It was also wanting to know more about the products themselves and their characteristics, and understanding the implications from those characteristics for how likely it is that we're going to see these products adopted and diffused across the country in the near term. Because the benefit is potentially so high, we really wanted to know how soon we might see that if we were finding technologies that were there.

Glen Woodbury

We also wanted to see that, even if a product is a capable product in other uses, that doesn't necessarily make it an operational capability for emergency management.

Deanna Lee

Okay, and we're going to get into some of these tools and products in a moment here. But before we do, I'm curious about who should be paying attention to these findings. I think emergency management might mean different things to different people. So when we say emergency management, who are we talking about here? State and local agencies, nonprofits, first responders? Is it all of those people?

Jessica Jensen

It is, it's all of the above. And I think the what of emergency management and the who go kind of hand in hand. When we talk about emergency managers, we are talking about people who work in government emergency management offices or business settings and have that title. They're the people who work behind the scenes to help organizations and communities get ready to help figure out how they're all going to work together in these really tough times—where the environment is urgent, the information is uncertain, and decisions need to be made to save lives, people, and property. So that's emergency managers. They're helping us figure out, behind the scenes, how to make the best decisions, how to divide resources in these circumstances, and before to get ready.

But the actual work of responding, if you think about what's going on in as a disaster is happening, right? You have first responders that are saving lives—not emergency managers. In the aftermath of disasters, you have nonprofits that are helping collect donations and help our most vulnerable in our communities. You have construction companies that are rebuilding homes, right? You have all kinds of different players that are involved. And so those folks are doing the work of emergency management too. And so the products that we wanted to identify are those that any of those folks could use. If you have a role in disasters, we wanted to know what was available to help you today. And that's a broad group.

Glen Woodbury

Yeah, and one way I think about it is that emergency management is not a single office or just a single profession. It's really a distributed system or a network. So emergency managers help create the framework and connect the pieces. But as Jessica said, the work itself is carried out across government, the private sector, nonprofits, and larger communities. So these findings matter not only to emergency management agencies, they matter to every organization that has a role in preparedness, response, recovery, or mitigation.

Deanna Lee

OK, so maybe you can take us inside the experiences of these professionals for a moment. I'm wondering what are the biggest pressures and day-to-day problems they might be facing that AI could potentially help with. I'm sure the list is long, but maybe you can just give us some of the highlights.

Glen Woodbury

Sure. Let's start off with the basic problem. It’s that demand on emergency management is growing faster than capacity is. Emergency management has long been under-resourced, especially at the state, local, tribal, and territorial levels. And that gap is widening. Agencies are dealing with more frequent and severe events, more hazards, higher public expectations, and responsibilities that were not traditionally assigned to emergency management in the past.

Events are also increasingly overlapping, and they cascade. For example, a wildfire can become an evacuation, a public health, a housing, a utility, and economic and recovery problem all at the same time. Emergency managers are also flooded with data, reports, imagery, requests, and a lot of administrative work. So AI may help them sort, synthesize, and act on that information faster. The value proposition is not replacing emergency managers, it's giving a limited workforce more capacity while preserving human judgment and accountability.

Deanna Lee

And are emergency management organizations already using AI in meaningful ways? I'm sure ChatGPT and other LLMs have found their way into their jobs, much like they have for a lot of our jobs. But did you find anything else in terms of how they're already using some of these tools?

Jessica Jensen

So we have both authoritative evidence and anecdotal evidence. So the most recent authoritative evidence we have comes from Argonne National Laboratory and the study, the results of which they produced just last year. It was nationwide—every level of government looking at the extent to which a technology was being adopted, including AI, and what the barriers to that were. That available information in 2025 suggested that emergency management is at its very nascent stages, with very, very few having access to or using AI-enabled products.

That said, a year later, the landscape of AI in this country is very different, and the available anecdotal evidence that we have suggests that they are very much experimenting with products on the market within the boundaries of what their organizations allow. LLMs for sure are getting a lot of attention. LLMs are large language models like you referred to—ChatGPT would be one of those. Claude, Gemini, there are others.

Glen Woodbury

As Jessica said, the adoption is uneven, and they don't always recognize that they're actually using AI already without calling it that. For example, forecasting, mapping, imagery analysis, translation, information monitoring are all using elements of AI-enabled tools, but they may not know that they are actually embedded in those tools that they already using.

Jessica Jensen

Yeah, if I could just take a second to actually say what we mean when we use the phrase artificial intelligence. First of all, there's no consensus-backed definition. Like, among the geeks like us, still having debates, right? Nobody has come to consensus on exactly how we're going to describe this thing consistently so that people understand what we're talking about. However, what we are generally talking about in the scope of this study is computational systems that take on tasks that would require intelligence—human intelligence—normally to do. So things like communication or reasoning or prediction or planning—those are the kinds of things that computational systems can do when equipped with artificial intelligence. And so that is what we're really talking about is those kinds of things. It's a big, huge range of actual techniques and tools that fall under that umbrella from things like Cloud Vision, which is using imagery and having artificial intelligence computer systems looking at imagery and achieving various kinds of analysis, and all the way to those LLMs we were just referring to.

Deanna Lee

Okay, right. Excellent. Thank you so much for clarifying that. And you actually identified more than 1,100 AI-enabled products that could support emergency management. So before we cover those tools, let's talk about how you found them, because you actually used AI to identify them, right?

Jessica Jensen

Yeah, so it was a crazy journey doing this study and we never would have predicted at the outset where we'd find ourselves. When we started out, to be honest, on the research team, we didn't have high expectations that we were gonna find already a lot of AI-enabled products on the marketplace, you know, maybe dozens. That's what we kind of assumed at the outset. And as we began to kind of scan the marketplace, talk to experts, to do independent searches of various sorts, we were overwhelmed by the number—the sheer number that we were finding—and the breadth of the kinds of ... ways they could be used and tasks they could support. It was incredible. So we never would have thought that we were gonna find over 1,000 of them.

Then, because we always had that second interest, right, we wanted to understand their characteristics. We wanted to know about the products themselves. So we wanted to understand, if based on those characteristics, they could be adopted, and we could see these diffused across the country. Well, that became an enormous problem at scale, right? Like how do you collect systematic, consistent information along a lot of different dimensions about 1,179 products? Well, I've never been more glad to be at RAND. Well, that's not true. I'm always glad to be at RAND. But, at RAND we have access to some of the most incredible geniuses on the planet—I am so lucky to call them colleagues—who were able to devise, using artificial intelligence, how to collect information at scale, essentially writing tons and tons of code that I know nothing about—tons and tons of the code, connecting it up to LLMs, different ones, and searching for that information at scale.

We had human verification processes that were really robust, really intense. We had automated verification processes, so a lot of checks and systems in place to make sure that the data we had was the very best available. But without AI, we could not have done this study at scale and in the timeframe that we had. So it was really cool to be part of that and to draw upon the expertise that we have here at RAND.

Deanna Lee

Yeah, incredible. And at a high level, I mean, 1,100 tools, we're not going to be able to talk about all of them. But at a higher level, what did you find? Can you give us an idea of, you mentioned Cloud Vision earlier, we've already talked a little bit about LLMs. But can you give an idea of some of the tools that were out there and maybe hit some of highlights of what you identified?

Glen Woodbury

Sure, and as Jessica and you emphasized, the first finding was just the sheer volume and the variety of tools. But we also discovered they're not evenly distributed across the emergency management phases. There are a lot more products out there that support the response phase, fewer for recovery, and many, many fewer for the preparedness phase. So the most crowded area within response was situational awareness, hazard monitoring, and information sharing. These tools can combine, for example, satellite imagery, drones, cameras, sensors, weather data. News and social media to help operators understand what is changing.

The second example would be damage assessment, generally under either initial damage assessment during response or more intensive damage assessments for recovery. So computer vision can compare pre- and post-event imagery, flag likely damage, and help teams decide where field verification is most urgent.

The third area was information management: AI can summarize situational reports, extract key details from documents and messages, translate information, and help organize large volumes of incoming data. There are also tools for forecasting, evacuation, transportation analysis, logistics, resource tracking, public communication and other parts of the mission.

So, for emergency managers, the challenge is not simply finding an AI tool, it is identifying a tool that solves a specific mission problem and can work within the agency's existing systems and constraints. So, for example, during a fast moving wildfire, there might be a tool that can combine fire spread, weather, traffic, camera, and public reporting data to flag a changing evacuation problem. The emergency manager still makes the decision. The tool helps surface that information sooner.

Jessica Jensen

The products we found, too—just to add on to what Glen is sharing with you—are available to support more than just those emergency managers, as we said, who work in emergency management offices and help us behind the scenes. So Glen's examples are excellent of the kinds of things on the market to help that audience. But those other players, right? Those nonprofits that help us in response and recovery, serve the most vulnerable, help us rebuild our communities—we found quite a few things available on the market that are AI-enabled that could either enhance their coordination and collaboration within the nonprofit community to benefit survivors, getting them more holistic help faster. We found products that could help manage volunteers at scale and across jurisdictions, again, enabling this inter-organizational coordination to help coordinate the efforts of volunteers across many different organizations. And then donations management. We found things that could help them do that.

Utilities are huge, huge players, right, in helping us have a good quality of life and functioning communities. We found all kinds of products on the market that can support water management, that can help support electricity, communication, infrastructure. It was incredible, the diversity of products that are out there that could help those organizations that support disaster work do their day-to-day work better and faster, and, in disasters, pay real dividends in saving people's lives and property, speeding along recovery.

Deanna Lee

Were there any tools that were particularly surprising to you? Did anything jump out as something maybe you wouldn't expect to have seen that you did discover in this study?

Glen Woodbury

Maybe not so much the specific tools, again, back to just the volume of how much already exists. And so it's less an issue of the absence of technology than the difficulty of seeing the market clearly and knowing what is credible, relevant, and sustainable. I think that was kind of, not necessarily surprising, but (there was) kind of more clarity through the research. I was also struck by the imbalance—that the market is heavily oriented towards the response and organizational needs. But with that less attention to preparedness, long-term recovery, and the experience of survivors. And frankly, gaps actually create opportunities, especially for the technology sector, to step into where those gaps are and help fulfill those needs.

Deanna Lee

That's a good segue. I want to ask you about major gaps that you identified. You know, we talked about the sort of breadth and depth of the tools that were out there, but were there areas that emergency managers have real needs but that currently aren't being met or filled by any AI tools that are out there on the market?

Glen Woodbury

We actually have an entire appendix on the gaps that we saw. So for example, some of them within the response phase would be those that serve fatality management services, emergency debris clearance and route clearance, family reunification, those types of things. Under recovery, natural and cultural resource restorations, private sector, public-private sector coordination and collaborations. A lot of things that bring discussions together seem to be gaps. Also, some gaps in land use, rebuilding, mitigation planning, those types of things were gaps that we found. And those are, like I said before, those are just some of them. We found quite a few and we documented those in the report.

Jessica Jensen

I think a big, huge challenge—it's more a challenge than a gap: a lot of the solutions that we found on the market are very, very narrow. Glen was talking about situational awareness as this really important thing that emergency managers are trying to accomplish. They're trying to get an understanding from a host of different data sources, right, of what's going on, where is help needed? Where does it, where's it most urgent? How do we get there faster? Who's involved and what are they doing, right? They're trying to understand all of this stuff from all these different sources.

Well, we found lots of tools that provide a portion of insight about a thing that's happening, right, but it does not connect all of those different information streams and data streams together, that's number one. So each one's doing this very narrow little slice—and not only does it not connect the data, put it all in one place—but it also doesn't synthesize it to provide them any kind of initial insights that might be valuable as they make their decisions. And so we call that ... it leads to what we call a stacking problem. So people would have to have a whole bunch of these and connect them. And that has all kinds of implications for buying them, how much they cost right away, and then to maintain them. You have to have different levels of knowledge to operate and avail yourself of these different tools. And so, you know, that's not so much a gap, right, as it is a big challenge related to those products on the market.

Deanna Lee

That's exactly what I wanted to talk about next, which is the sort of implementation challenges. And it sounds like a stacking problem is exactly that. You also mentioned costs. Are there any other challenges that are kind of in the way of emergency managers taking these tools from this list that we have and actually implementing them in their daily work in ways that can help them?

Glen Woodbury

Sure, some of the things we found was that, you know, the biggest barrier may not be product availability. I mean, some is—to what Jessica just mentioned about the stacking and the integration of various tools. But it's also the organizational capacity to evaluate, acquire, integrate, govern, and sustain a product after its initial demonstration. Emergency management agencies already operate with long-term resource constraints and tend to be appropriately cautious about new technology, especially in high consequence situations. At the product level, agencies face dependencies on data and other platforms, integration requirements, connectivity needs, opaque pricing, uneven vendor support, different levels of cyber, privacy, legal, and operational risks.

Emergency managers, I mean, their job is risk management, which tends to also make them a bit risk averse. So they have to walk through all these governance and data and integration issues before they can fully integrate a product. So, the issues are manageable one at a time, but taken together, they can overwhelm, especially small offices that don't really have dedicated technology, legal or data staff. And that means larger and better resourced organizations may be able to adopt sooner, while smaller jurisdictions risk being left behind unless there's some type of shared support.

Jessica Jensen

So, getting into just a couple of specific examples, most of the products on the market require technical expertise, someone who knows IT, right? It requires their involvement initially and/or on an ongoing basis in order to onboard and/or use that tool. Not every organization is going to have that kind of staffing available, certainly at the outset and on an ongoing basis combined. But it could be either one, and that would be a problem if you don't have that support available to you. Additionally, most of the products, nearly all require internet connection—consistent internet connection—to use. Now that might not be a a problem if you want a product for preparedness, right, that you plan to use on a day-to-day basis. But if you're planning to use it in response, that can become a really big issue, and that was a massive trend we saw on the market.

We saw on the market that finding out what the price was of these things was impossible. And I can conclude that really, really on solid evidence. We had huge teams of those geniuses I was referring to earlier, trying to figure out how to get that information and at a level that we could trust. It was not possible for us to do that. Well, if you don't know the price of something, then how do you know if you can purchase that product and use it on an ongoing basis. And compounding this issue of not being able to tell how much things cost is the fact that we discovered a lot of the products on the market require another product to even function. You may only want the add-on functionality you'd have to pay extra for, but you would have to have the base platform as well. And that's true in a significant proportion of these products.

So, again, as Glen was saying, when you start to stack together the trends that we were seeing about these products and you couple them with ongoing risk concerns like privacy—most products have, there's some sort of privacy concern evidenced in the product, what it needs or how it operates in terms of really protecting information that's of a personal nature, like personally identifiable information, or of a jurisdiction or organizational nature. So, right out of the gate, you have a huge risk concern with most of those products. And then a lot of the companies out on the market don't provide outward facing evidence of cyber certification of any kind. So you couple privacy concerns with a product with the fact that we're not sure, we can't tell if you're really protecting that information. And now all of a sudden, a lot of these products that look so super sexy on the surface—like, oh my gosh, we could be saving bazillions of lives—it seems like really unattainable for a lot of organizations to be able to onboard and use these kinds of products on an ongoing basis.

Deanna Lee

A lot of obstacles, to be sure. And maybe even at a more basic level, I'm wondering how are emergency managers and organizations finding out about these tools? If they are, you know, maybe they aren't, but are these on their radar, even?

Jessica Jensen

So this is a really big challenge. It was, as mentioned, we had to use tons of different approaches to even find the products, identify that they exist, much less than to also gather information about them. There isn't a central repository or clearinghouse that exists for the emergency managers who are doing the work, much less the broader organizations that are involved in disaster work. There isn't somewhere to go. There isn't standardized information about the products on websites, vendor to vendor, or in some central place. So it is very, very difficult for them to do that.

That issue is compounded when you consider how vendors describe their products. Sometimes they're using really technical words. If you don't have that technical background, you don't know the implications of those technical words for you and your organization. And that's just a start of some of the kinds of challenges that emergency managers face as they try to learn about those products. It becomes a really laborious process, and often relies on word of mouth, like who around them is using what, and vendors that actually are proactive and approach them, and some combination thereof. But Glen, you probably have some additional examples of how they learn about it.

Glen Woodbury

Yeah, exactly. Emergency managers, as Jessica said, rely heavily on trusted peers. So some sort of credible clearinghouse, practitioner networks, independent evaluations, maybe even shared procurement support can make the market much easier to navigate for emergency managers. And vendor claims alone are not enough, as Jessica mentioned. Agencies need evidence, comparable information, and feedback from organizations that have used a product under realistic conditions.

The emergency management community is actually fairly small. And that's a huge benefit, but it also means that, as tools and products and other technologies come into their space, they talk to each other and they wanna know what's working and what's not working, and they'll share that information. But that could be a bit more rigorous and a bit more intentional to really get AI-enabled tools out there to the community.

Jessica Jensen

So I would add, because we're talking about an audience that's bigger than just those government emergency managers, that the research actually says—not our research, but long-standing research about technology diffusion—what are the things we see when technology becomes diffused? This whole peer network thing's a big deal. It's a big deal, but it also means when we're trying to solve disaster problems, it's a big deal for every field. So what we would want to see is that all these different players in disasters in the respective field have these peer leaders and advocates, have those networks of learning, have that kind of exposure to those products to learn about them and to diffuse them. They too face the problem of no centralized clearing house with respect to the products that might serve them. Now that may evolve in any number of these spaces over time, but our question was, what can we expect in this near future? And in the near future without greater and more significant mechanisms to propel this diffusion across these fields, we can't expect to see a ton of progress in the near term.

Deanna Lee

Okay, let's set the near-term aside for a moment. We've been talking about these challenges that need to be addressed if AI is going to live up to its promise in terms of how it can transform emergency management. And I want to ask you about that promise. If emergency managers can successfully implement these tools in their work, how might it transform the sector?

Jessica Jensen

Well, maybe I'll start with preparedness. And then, Glen, if you want to talk about response or recovery or something. Preparedness is that time before anything happens. In the time before everything happens, the reality is that a lot of time is consumed with administrative tasks in these offices across the country. Maybe they're seeking grants or reporting on grants, they're preparing plans, they're developing—we call them exercises, activities that help organizations practice, how they're going to respond. It's doing training. It's developing public outreach materials. Emergency managers spend a ton of time on those activities. Tons of time. What AI can do is alleviate the amount of time that is spent on those tasks. It can also save a lot of money. Emergency managers often contract out some of those tasks, so there's expense involved as well.

So AI can alleviate so much of the burden of the normal routine preparedness work and allow emergency managers to be talking to people like us, to be engaging one-on-one in communities, going to schools, advocating for recovery planning, advocating for ways to reduce the likelihood that events are gonna even happen in their jurisdiction. They can shift their focus to be more engaged with people in their communities and organizations.

Glen Woodbury

And I'll start with the response. So the response phase is where lives are saved and property is protected. It's that actual kinetic phase where these things are happening. And as we mentioned before, the emergency management field is generally doing the coordinating and connecting, while the first responders are out there doing those. But the clearest opportunity from AI is faster sensemaking. So AI can pull together multiple information streams, flag changes or anomalies and reduce the time that staff spend manually sorting information to get that to the first responders to do that life-saving mission. For example, situational awareness tools can help maintain a more current operating picture. Damage assessment tools can rapidly screen large areas and prioritize where people need to go to either save lives or to inspect damaged property to see if more work needs to be done.

Also during recovery, AI can help organize imagery, applications, reports, and other documentation to identify missing or inconsistent information. It can reduce repetitive administrative work in the recovery phase. I think the goal is to shorten that distance between information gathering and sound decisionmaking. And it's not to automate it, but it's to make better sense of what is actually occurring out there in the field—to speed decisionmaking so that we close that gap between lives that need to be saved and lives that are saved.

Jessica Jensen

If I could add, Glen, you know, it could also really help us predict and anticipate where needs are going to be. And this becomes a really particularly acute opportunity based on really long-standing research that suggests that we typically see disproportionate impacts of disasters within communities. For example, the disabled suffer more and their recoveries are longer than other segments of the population. With the benefit of AI and the benefit of all these different data sources coming together like Glen is describing, we'll be able to predict and anticipate where the needs are gonna be greatest, where our most vulnerable are at, and get them the kinds of help they need in appropriate ways at the right time. And that has been a historical challenge that's faced this field. And those are very real human outcomes that can be improved upon going forward with the benefit of AI.

Glen Woodbury

Yeah, and I'll add that this idea of sensemaking, one of the biggest pieces of that are recognizing weak signals. So human beings, when the adrenaline's running and you're running the EOC and the operations and the lights and sirens, you pay a lot of attention to the big noisy things, but you might be missing some of the weaker signals that will give you more information. AI can kind of remove that adrenaline rush that's right in front of you, looking at the big issues, and help identify those things that may be around the corner, over the horizon, or that you're just not seeing along with all the noise that is occurring during the response. I think AI could help with that piece of the sensemaking as well.

Deanna Lee

Absolutely. It sounds like there's probably a long way to go before we sort of realize this future you both have just described, but what's your advice for emergency managers and policymakers now? What should they be thinking about as they consider or start to consider how they might use AI for the important work that they do?

Glen Woodbury

Well, let's separate emergency managers and policymakers because they're going to have different requirements and different things that we could give them advice on how to approach this. I'll start with the emergency managers, and Jessica can take on the policymakers. For emergency managers, I would start with the mission problem, not with the technology. So be clear about what you're trying to improve and know how you're going to evaluate that tool and whether it actually helps. So we talk about this in the report, that we need to assess organizational readiness in an emergency management organization across several areas: applicability, adoptability, adaptability, integration and dependencies, reliability, governance and risk, procurement and resources. And can the organization train and learn and measure performance, learn from use, and stop or change course when needed? So tested exercises, routine operations, involve frontline users, establish human review—it's more than just the purchase. An organization has to go through quite a few other checkpoints or at least evaluate the tool in several different areas before they can make it adoptable to that organization.

Jessica Jensen

Thinking just for a second about emergency managers, too, is technical knowledge and expertise is not the reason we hire a lot of people to help us with disaster stuff. And so for a long time, research has documented that those that serve disasters, whether they're in emergency management offices or otherwise, do tend to lack some of that foundational technical expertise to even do that evaluation that Glen is talking about. So a big opportunity for emergency managers is to seek professional development in the area of AI and understanding it—just becoming conversant about the different techniques, how it shows up in different products. They could start with our report, but there are tons of other resources out there that they can avail themselves of. And I know just, hint, hint that the Markle Foundation will be funding some efforts to help upskill the workforce as well as the initiative moves forward.

Talking about policymakers for a second, this is really big for me. I'm going to talk about policymakers, but another audience, technology companies, the ones that are making the products. One of the most confounding things about our research is all the problems—oh my gosh, all these challenges. One of the most beautiful things about the work that we did and what we found is there are so many opportunities to fix it and players who have control. Relative to policymakers and technology companies, emergency managers have very little control here. They work within jurisdictional and organizational environments where IT departments and those leaders make a lot of decisions about how things are, what's allowed, what the cyber risk is, how risk averse are we to it. That's a huge, huge part of this—whether they can even adopt the tools in the first place. Then you have other entities that control procurement processes, how we buy products, what are the rules around how we buy products, how long does it take us to buy products? That is another huge factor here and policymakers control that. Then, you have those elected officials that make budgetary decisions—who gets what money for what? And as Glen said, emergency managers historically have not had enough resources to do the very foundational life-saving work that is their mission. So, going forward, policymakers can choose to invest in the lives and welfare of those populations they serve by investing more in emergency management so it can get the tools it needs to do the work better and faster and at scale than it ever has before. So, so much potential there.

Also, just to put it out there, AI governance, this whole thing of what is AI, what are we gonna even allow? Those kind of decisions are being made across the country right now. It is too early to say whether those will facilitate emergency management's ability to avail themselves of these tools or not, but have no doubt the decisions they make are going to control what happens. And so before I go to emergency management people and say, hey, start adopting tools because they're going to help you. I'm going to say to the policymakers, hey, what are you doing that is going to facilitate or pose a barrier to their ability to avail themselves of these technologies.

Looking over on the technology company side, wow, so much opportunity, this is in general, a low capacity field; design your products with that in mind. It's generally a low technology knowledge community; communicate about your products in ways that lay people can understand. Price is a huge issue; don't hide it. You know, there are very, very basic things that technology companies can do if they're just cognizant of the way the characteristics of their products interact with these things that either facilitate or create a barrier to people buying their products.

So this is really good news, I think. I mean, I see such potential for these audiences to contribute in meaningful ways. And, make no mistake, like disasters are happening faster and faster and fast, they're overlapping, they're of different kinds, they're affecting more people, they're more and more expensive. We cannot afford the trajectory we're on. Like we cannot sustain this going forward. Change has to be made. And emergency management is the space where we make those changes. So those investments are important. It's important to you, it's important to me. It's not important to the emergency manager as much as it is to us. And so those policymakers, those technology companies, it's in all of their interest to take a look at what actions they can take so that these technologies can benefit all of us.

Glen Woodbury

And I think I'd add on for the technology companies, better understanding of the mission space that emergency management and emergency managers work within—as opposed to letting the tool create the requirement, the tool should serve the mission. And that would require those that are developing these tools to understand that mission and the tasks within it. I've been around for some time. I've seen many technologies that have been adopted that created more requirements that didn't serve the mission, and more staff time, and more labor, and more cost as opposed to actually alleviating any of those things.

Deanna Lee

Excellent points, both of you. We've covered a lot of ground today. So before we go, I want to hear from both of you, what's the one thing you want listeners to take away from this discussion they've heard today?

Glen Woodbury

I came away both encouraged and cautious. There are far more potentially useful tools than many emergency managers realize are out there. But the market is difficult to navigate, and important gaps and challenges remain. I think the opportunity is to match technology to the real mission needs, designed more deliberately for underserved parts of the disaster lifecycle and for survivors, and adopt tools in a way that manages risk rather than ignoring it. I think AI can expand, certainly, capacity and improve sensemaking. But it should support, not replace, professional judgment, public accountability, and the relationships that make emergency management work in the first place.

Jessica Jensen

Glen, that was awesome. So how do you even follow up with that? Also, it's like 10 things, not one thing. But that aside, I think I really want people to understand that we, we don't need tons more technologies. That's happening. It's there. It's there now. We need to pay attention to the implementation environment, because that is equally, if not more, important than the existence of these tools themselves. That is essential. That is essential for everyone to understand as you're learning about those products, learn about the environment they need to be implemented in, and from whatever space you're in, technology companies like Glen mentioned, policymakers, emergency managers, all of us need to be aware of both sides.

Deanna Lee

Okay, I think that's all the time we have for today. Jessica, thanks so much for joining us.

Jessica Jensen

A pleasure. Thank you for having us.

Deanna Lee

And Glen, we appreciate you being here as well.

Glen Woodbury

Great, thank you, Deanna. I appreciate it.

Deanna Lee

And of course, thank you to our listeners. If you want to read more about the report we discussed today, you can find the links at rand.org/policyminded. You can also learn more about the Markle Foundation's AIDE Initiative at aidinitiative.org, that's A-I-D-E initiative.org.

Today's episode was recorded and edited by Emily Ashenfelter. It was produced by me, Deanna Lee. RAND's Director of Digital Communications is Pete Wilmoth.

RAND is a nonprofit institution that helps improve policy and decisionmaking through research and analysis.