Emergency Management and Generative AI
By Dan Stoneking, writing for Homeland Security Today
It seems like generative AI has reached the venerable level of the Oxford Comma debate. I actually published a strategic communications column about the Oxford Comma a few years ago where I definitively proved that being for or against was missing the point. It is actually a matter of choosing when, where and how to use it (or not) consciously and strategically.
And now it is time to provide that same analysis for generative AI. Who could have predicted a few years ago that we would all be so invested in debating and investigating the em dash?
Before I dig in, let’s consider some historical perspective. When the printing press was invented there was fear that everyone would stop going to the theater and scribes would cease to exist. That did not happen. Then the radio was supposed to replace reading, television would eliminate radio, and electronic devices would eradicate books.
None of that happened.
Fast forward to the last five years. We seem fine with the dictionary, the thesaurus, digital cameras, and internet search tools, but once again, we fear the latest tool.
Like the printing press, the newest tool is here. Siding for or against is a moot point. The better question is how do we use it. We need to address the do’s and don’ts.
Overview of Generative AI
Generative artificial intelligence is a class of tools that produces new content including text, images, audio, video, and even code. These systems are built on patterns learned from massive datasets. The process is simple in concept. Give it a prompt, and it produces something that increasingly looks and feels human. It will never give you gobbledygook (nor commonly use that word)
In emergency management and adjacent fields, these tools are already being used for drafting plans, summarizing long documents, building training materials, translating messaging, and accelerating research. Outside EM, they are everywhere: marketing, education, software development, customer service, and creative work. The pitch is always the same. Faster. Cheaper. Easier. See my last three sentences. Generative AI does not typically favor one-word sentences, but I still do.
But none of this is magic, and it definitely is not neutral in impact. Generative AI is not good or bad in itself. It reflects how we choose to use it. That includes the part people often skip over, the environmental and economic cost. These systems require large-scale computing infrastructure, which means energy consumption and water use for cooling data centers. Image and video generation are especially resource-heavy, but even text at global scale adds up quickly.
This is not an argument to panic or abandon the tools. It is a reminder that “free” output is never actually free. And to put that in perspective, global AI data centers are projected to use as much as eight hundred billion liters of water annually. That sounds staggering. Until we see reports that chemical manufacturing uses two to five trillion liters annually. And the global beef industry uses several trillion liters annually. Not defending AI. Just providing context.
If we are bothered by one, we should be bothered by all.
For the rest of this piece, including Dan’s list of Do’s and Don’ts, click here.