Three Stories That Should Change How Fire Chiefs Think About AI Right Now

 

Three Stories That Should Change How Fire Chiefs Think About AI Right Now

By Ken Lord | AI Fire Guy | aifireguy.com

 

Last week dropped three AI stories in rapid succession. On the surface they look like tech industry and Wall Street news. None of them mention the fire service. But if you lead a fire department and you're trying to figure out where AI fits, or whether it fits at all, these three stories are more relevant to your department than anything else that hit the wire last week.

Let me walk through each one and then tell you exactly what I think it means for you.

 

Story #1: Anthropic Files for an IPO

On June 1, 2026, Anthropic, the company that builds Claude, confidentially filed a draft registration statement on Form S-1 with the SEC for a proposed initial public offering of its common stock.

The filing follows a $65 billion funding round that valued the company at roughly $965 billion, and a public debut above the $1 trillion mark is now considered the base case by many investment bankers. Around 80% of its revenue already comes from enterprise customers, with Anthropic's share of the enterprise AI market surpassing OpenAI's for the first time in April 2026.

That's a lot of numbers. Here's the one that matters to you: 80% enterprise.

When a company goes public, it answers to shareholders and quarterly earnings calls. That changes everything about how it prices its products, who it prioritizes, and what customer support looks like. The startup flexibility, straightforward pricing, simple sign-up, manageable terms, doesn't survive contact with Wall Street indefinitely.

I watched this exact pattern play out in the fire service with learning management systems. When I was a captain, our LMS contract was a handshake deal. The vendor was local, the pricing was straightforward, and you could actually get someone on the phone when something broke. Then private equity rolled in, the companies got acquired, and suddenly every department in the county was navigating multi-year enterprise contracts with eight-month procurement processes and pricing that bore no relationship to what small departments could actually absorb. The product didn't change. The ownership structure did. That was enough.

The departments building AI competency right now are doing it at startup prices with startup flexibility. That equation is starting to shift. The practical risk isn't that your $20/month Claude account is going to disappear. It's that serious departmental integration, the kind where AI connects into your CAD, your RMS, your training management system, will only be available through enterprise contracts with procurement requirements, compliance layers, and price points that reflect a public company's obligation to shareholders. The departments that build that competency now, while the tools are accessible and the barriers are low, will be significantly ahead when that shift completes.

 

This week's action: Find out what AI tools anyone in your department is currently using. Write them down. If you can't name them without asking around, that's your answer about where you stand.

 

Story #2: Trump Signs an AI Executive Order

On June 2, President Trump signed an executive order titled 'Promoting Advanced Artificial Intelligence Innovation and Security,' marking a shift by the administration toward federal oversight of AI.

The order establishes a framework for the federal government to vet the national security risks of the most advanced AI systems for up to a month before their public release. Participation by AI developers is voluntary. It also directs federal agencies to develop benchmarks to assess AI models' cyber capabilities and create an AI cybersecurity clearinghouse to review and share information on vulnerabilities.

Voluntary today. That's the operative word.

Here's what 40 years in the fire service taught me about voluntary federal frameworks: they become mandatory grant conditions. Every time. I watched it happen with cybersecurity requirements in BSIR and AFG applications. A few years ago, cybersecurity was a recommended best practice in federal grant guidance. Then it became a scored element. Then it became a checkbox you couldn't skip. The departments that had already built their cybersecurity posture didn't scramble. The ones that hadn't found themselves writing policies in a hurry to stay eligible.

AI is at the same early stage of that exact cycle right now.

Think back to when social media arrived in the fire service. For a few years nobody had a policy. Then a firefighter posted something from a scene, or made a comment that blew up, and suddenly every department in the county was writing a social media policy in crisis mode. Most of those policies were reactive, poorly thought through, and created more confusion than they resolved. The incident that forces a reactive AI policy hasn't happened to your department yet. That doesn't mean it isn't coming.

The federal government has also already moved past the discussion phase in some areas. In May 2026, the Department of War announced agreements with eight leading AI companies to deploy their capabilities on classified networks. The military is past the 'should we use AI' question. Federal grant programs follow that lead, and they always bring compliance strings with them.

 

This week's action: Pull out your department's acceptable use policy. Does it mention AI tools at all? If not, that's a gap you can close before someone closes it for you in a way you don't control.

 

Story #3: Uber Burned Through Its Entire AI Budget in Four Months

This is the most instructive story of the three for department leaders, and the one getting the least attention in fire service circles.

Uber exhausted its entire planned 2026 AI coding budget in just the first four months of the year due to soaring usage. Individual software engineers were generating bills ranging from $500 to $2,000 per month. The company has since capped monthly employee spending at $1,500 per tool, targeting agentic coding software like Cursor and Anthropic's Claude Code.

Uber, a company with billions in annual revenue and a full technical staff, didn't see that coming. They set a budget in 2025, before anyone fully understood how fast usage would accelerate, and were tapped out before spring.

Now translate that to a fire department context.

Your training captain starts using Claude daily to build lesson plans, update SOGs, and draft after-action reports. Your PIO starts using it for press releases and social posts. Three battalion chiefs start using it for shift reports and pre-incident planning. None of them are coordinating. None of them know what the others are spending. You've got a handful of people running two or three tools each across the department, and nobody has visibility into the aggregate cost or whether any of it is producing measurable results.

That's not a disaster scenario. It's actually a pretty reasonable picture of how AI adoption starts in most organizations. The problem is what happens at the next budget cycle. When your city manager or county administrator asks what you're spending on AI and what you're getting for it, you have no answer. And in the current fiscal environment, 'I'm not sure' is not a position you want to be in when a city council is looking for cuts.

I know that many departments reading this are running tight budgets where adding another officer is off the table, let alone new technology spending. That's exactly why this matters more for smaller departments, not less. When you can't absorb a budget surprise, you need to see it coming. The departments that manage AI spending well are the ones that will be able to defend and expand their AI programs. The ones that can't answer basic ROI questions will get the plug pulled.

Uber's response was practical: they deployed an internal dashboard so staff could track real-time costs, established a review process for employees who needed to exceed limits, and set per-tool monthly spending caps. That's not bureaucracy. That's the same operational accountability we apply to apparatus maintenance, overtime management, and supply budgets. It's how you run things responsibly.

Also worth noting: despite executives reporting that AI generated at least 10% of Uber's code, the new technology still appeared to be hurting the bottom line. Adoption without measurement is just spending. The ROI question is coming for every department that introduces AI tools, and the ones that can answer it clearly are the ones that survive the first budget cycle where someone decides AI is an easy target.

 

This week's action: If your department is using any AI tools right now, write down one specific thing it produced, with a time savings or outcome attached. A training outline that took 20 minutes instead of two hours. An SOG section drafted in an afternoon instead of a week. If you can't produce one example, you have a measurement problem, not an AI problem.

 

The Pattern All Three Stories Share

The common thread across all three stories is this: AI is entering its enterprise accountability phase.

The 'figure it out as you go' period is not completely over, but it's closing. The companies building the tools are going public. The federal government is building oversight frameworks. And well-resourced corporations are discovering that unmanaged AI adoption creates financial problems nobody planned for.

For the fire service, this means the window to build a smart, simple AI framework on your own terms is still open, but it won't stay open indefinitely. The departments building that foundation now get to do it proactively, on their schedule, at their pace. The departments that wait will build it in response to a grant requirement, a budget question, or an incident that forces the issue.

You don't need a 40-page policy document. You need answers to four questions: What AI tools are we using? Who is authorized to use them? What data can go into them and what can't? And what are we actually getting for it?

That's the starting point. Everything else builds from there.

 

Forty years in this job taught me that the departments that handle change well aren't the ones who waited until everything was figured out. They're the ones who made a reasonable decision with the information available and adjusted as they learned. The information is available. The decision is yours.

But understand: not deciding is still a decision. And right now, it's one that costs you position.

 

Before you close this tab: Write down every AI tool anyone in your department is currently using. ChatGPT, Claude, Copilot, anything. If you can do it in 60 seconds without asking anyone, you're ahead of most departments. If you can't, that's the most important thing you learned today. Drop what you found in the comments below. I want to know where departments actually are right now.

 

Ken Lord is a retired  Battalion Chief with 40 years of fire service experience. He runs AI Fire Guy at aifireguy.com, where he helps fire service leaders understand and implement AI in their departments.