Playing with Fire, Safely: How Operations Leaders Can Light the Way on AI

Kevin O'Marah

September 24, 2026 • 4 min read

---

News that the US and China are prepared to talk about a new AI safety notification mechanism probably sounds like a good first step to people worried about the dangers of AI, including job destruction, rogue agents, cybersecurity, and social instability. 

If AI is an existential threat but also a race no one can afford to lose, then who builds the guardrails? Political leaders appear paralyzed in a strategic trap while investors try to time the investment bubble, and technologists plead for regulation. Safety sounds great, but what does it mean when we have no choice but to learn by doing?

Maybe it’s a job for operations leaders who approach safety in practical, human-centric terms but still optimize for getting the work done.  

What If AI Is Like Fire? 

We often look to history for similar technology revolutions in the human story. Electricity, the internet, and nuclear power are examples, but none have happened as quickly, nor as pervasively, as AI has since 2022. A better comparison might be the mastery of fire 400,000 years ago in Britain. 

Fire, like AI, is easy to use and even easier to demonstrate. Its applications include light, heat, cooking, craftwork, and conflict. Archeologists note also that it was a “catalyst for social and behavioral evolution,” enabling early humans to gather around the fire after dark, developing language, religion, and technology. And yet, fire also gets out of control as seen throughout history in London, Chicago, and, most recently, Pacific Palisades.  

Ask yourself, “What if we had never learned to work with fire?” 

Stopping AI Is Not an Option 

Bill Gates, in his recent letter on AI safety, starts with recognition that turmoil is inevitable: “The transition to the AI era will be one of the most turbulent times in human history.” AI may be widening gaps between rich and poor, right and left, old and young. It has also already shown the ability to escape human control and cause trouble. 

And yet nearly 1.6 billion people on Earth have used generative AI, with penetration levels above 40% in more than a dozen countries. More important for the flywheel of discovery we’ve kicked off, AI developer activity has increased 83% in less than a year.  

The train has left the station, and we are all on it.  

Safety First 

The fire analogy should resonate for supply chain and operations leaders. Many manufacturing and distribution processes involve the controlled use of fire, including smelting, welding, baking, and, of course, internal combustion. Human work around these processes is dangerous and normally governed by rules established over decades of trial and error. 

Safety is typically an obsession for site leaders in plants, DCs, and field service depots. Senior leaders often start meetings with a “safety minute” or some other kind of ritual focused on operator safety. Most also believe safety improves productivity. 

This mentality is exactly what AI needs: the willingness to handle fire, and the discipline to do so safely. 

How Operations Can Lead on Safe AI 

Getting practical involves at least three techniques that any organization can apply: 

1. Train people on the tools. Vibe coding tools like Lovable, Claude Code, and Replit allow users to build their own agents in minutes. Zero100 has hosted vibe coding workshops with dozens of companies, which build excitement about AI, but also understanding of what can go wrong. 

2. Form tech-ops fusion teams. Closing the understanding gap between Wizards (aka, technologists) and Citizens (everyday users) happens best when they work together connecting workflows across the business. Each helps the other learn what works and what hazards to watch out for.  

3. Establish an AI agent governance board. Agentic systems security is less about system access and more about agent behavior and adaptation in workflows. In partnership with the CISO, a governance board of ops, tech, and infosec leaders work together to safely scale agentic AI in the supply chain. 

The idea is to master AI by applying it conscientiously in practice rather than debating regulation from an ivory tower. Accidents are inevitable, but safety principles developed over decades in operations will keep us moving forward while building solid guardrails. 

Play with fire. Safely.