Let's talk about AMBER Alerts. Again. It really is the alert type that pops up as one of the most problematic with public message receivers in terms of messaging and policy. The messages tend to lack actionable details and the timing at which they are sent is frequently questioned. In fact, these two issues, combined with the geographical scope of the alert itself, appear to contribute to the majority of people opting out of AMBER Alerts on their phones. Because the prob

Should we use a link in our emergency alerts? Will people click them? Are some more trustworthy than others? We have research to answer each of these questions. In 2019, WEA 2.0 made it possible for alerting authorities to include a URL to guide people to additional information. This was a huge win for the emergency management community! Until then, we were limited to shorter messages AND no way to offer additional information. Then, everything changed and it was suddenl

Coded language does not simplify a message, it makes it more complex. When it comes to communicating danger, we need to know what the danger is, why its a danger, and what we can do about it.

AI-generated Alerts and Warnings? Part 7 Originally posted on LinkedIn on May 2, 2026 as part of a series on AI-generated alerts and warnings. Test bank – answer key – grading rubric. Now under construction. To be honest, it took me quite a while to understand what my EM1 colleagues were interested in doing with an AI-generated Alert and Warning benchmark. It took months of back and forth before I was assured that I understood what they needed for running tests, including t

AI-generated Alerts and Warnings? Part 6 Originally published on LinkedIn on April 25, 2026 as part of a series on AI generated Alerts and Warnings. I recently had a conversation with a team of researchers who commented on how transparent I’ve been with my articles about building a benchmark for AI-generated alerts and warnings. Being so open about the planning, the process, and the steps we are taking is unexpected, it seems. And yet when EM1 asked me to come on board as a

AI-generated Alerts and Warnings? Part 5 Originally published on LinkedIn on April 18 as part of a series on AI-generated Alerts and Warnings “If we can’t measure it, we can’t improve.” That’s what Tyler Felous, CEO of EM1 said about benchmarks. And agree with this 100%. Using a data-driven approach means measuring success and failure in order to identify ways to improve. But what exactly would a benchmark measure? Last week, I described the ways that medical researchers

AI-Generated Alerts and Warnings? Part 4 Originally published on LinkedIn on April 11, 2026 as part of a series on AI-generated Alerts and Warnings For the past few weeks, I’ve been writing about AI for Alert and Warning and the need for benchmarks that set standards that LLMs should adhere to. I’ve had a few research groups reach out to me about this and I’m happy to see that there is academic and practitioner interest in the project that I have begun with my colleagues. I

AI-generated Alerts and Warnings? Part 3 Originally published on LinkedIn on April 4 as part of a series on AI generated Alerts and Warnings. In a previous article, I introduced the idea of establishing benchmarks for AI Alert and Warnings (see this link). I explained that AI A&W benchmarks can help to set standards that will increase trustworthy use of AI while also ensuring that AI-generated messages align with the state-of-the-art in research and practice. I also introduc

AI -Generated Alerts and Warnings? Part 2 Originally published on LinkedIn on March 28, 2026 as part of a series on AI generated Alert and Warnings. Why are benchmarks valuable to Emergency Managers and other public safety organizations? Well, without them, you don't have a standardized way of determining if something is “good enough.” In medicine, researchers conduct clinical trials to determine the effectiveness of a new drug. In public health, researchers conduct experim

AI Generated Alerts and Warnings Part 1 Originally posted on LinkedIn March 21, 2026 as part of a series on AI Generated Alerts and Warnings I've been thinking for a while about AI generated alerts and warnings. As AI capabilities increase, it's a reasonable thing to consider. After all, I've seen students in my training classes load my published papers into ChatGPT and prompt it to write a message that follows the guidance I've written. And it's no secret that there are some

Organizational response information isn't recommended by alert and warning researchers, and yet it sneaks its way into messages here and there, especially those written for law enforcement activity.

Are you prepared to issue a message for an unusual threat such as a dangerous animal? Every now and then we find a WEA that is sent for an unfamiliar hazard, such as a bear in a populated area, an aggressive moose, or a cat or skunk with rabies. They represent the kind of message that is urgent with potentially severe consequences, but they probably aren't part of most organizations' message libraries. Should they be? I don't think it's possible to have ALL threats in yo

During the month of March, states throughout the U.S. recognized severe weather awareness week by conducting tornado drills. These included Virginia, Indiana, Kentucky, Michigan; in at least one state, the statewide tornado drill was cancelled (Missouri) and in other states these drills included local tests of the Wireless Emergency Alert system. We collected a bunch of the test messages and thought we’d go on a bit of a road show to see how they compare. Hop on board and l

Determining how to get a message to an evacuated population is a challenge; issuing a complete message can help to reduce message confusion.

A plain language missing person alert This is a really well-written alert. Let me show you have it conforms to the best guidance we have on how to write an actionable missing person alert. It begins with the source of the message: VBALERT. This is an acronym (it also appears like a branded alert name), however, at the end of the message we also see the organization clearly named. This is followed by the incident type (aka hazard from the Warning Response Model). The inci

An EBONY alert Recent research on effective Missing and Endangered Person (MEP) alerts has identified how to name the incident and how to structure a message to be most effective at helping message receivers to understand and respond to messages. When we consider MEP messages in relation to more general alerts and warnings, they follow the same structure as those warnings that are issued for all hazards. They should always contain the following CONTENTS: The name of the m

Required Weekly Test IPAWS users are required to regularly test their user proficiency by sending an alert that is issued to users who have opted in to receive them. This is also a great opportunity to demonstrate the type of test message that could be issued during a live test in a community - one that is announced ahead of time, organized with community members, and publicized by the media. These live tests have the added benefit, if done correctly, of growing system awar

Issuing alerts about warming centers should be easy to understand and accessible for all.

There are now more than 40 named alerts for missing and endangered persons across the United States. These messages are specific to individual states, and are commonly accompanied by state legislation to identify a vulnerable population that was not previously highlighted in missing persons messages. For instance, Arizona recently initiated the "Turquoise Alert" described this way : A Turquoise Alert program establishes a notification system that provides immediate informat

When we first began to study public responses to Wireless Emergency Alerts in 2012, we immediately saw significant issues with alerts that were limited to 90-characters. Study participants indicated that these messages could not contain enough information to act upon, so instead of leading them to protect themselves, it led to them to search for more information ( see Bean et al. 2016 ). Information search to confirm the message is trustworthy and requires action is commonl