Drop Your Tools - What We Can Learn About AI Adoption from Firefighting

This article is inspired by an article by Karl Weick.

The Disasters

In 1949, a crew of fire fighters was caught by a rapidly expanding fire in Mann Gulch, Montana. Their foreman, Wagner Dodge, ordered them to drop their heavy tools and run. He even invented an escape fire on the spot, a completely novel survival technique. Most of the crew ignored both instructions. They kept their tools, failed to outrun the fire, and 13 died within sight of safety. In 1994, nearly the identical thing happened at South Canyon, Colorado. Fourteen firefighters perished carrying chainsaws and packs they refused to shed, even as the fire overtook them. Post-incident analysis showed that dropping their tools just five minutes earlier would have given them enough speed to survive.

The haunting detail is the ordinariness of the failure. These were trained professionals. They kept their tools because those tools were their competence, their identity, their reason for being there. Without them, they didn't know what they were.

The Link to AI

AI is like the fire. It is not arriving gradually. It is already moving up the slope, and it is accelerating. Like the Mann Gulch blowup, it is changing the terrain (the world of work) faster than most people's mental models can track. The tools people have carried throughout their careers, specific technical skills, established workflows, professional credentials, ways of reasoning and producing, are increasingly being matched or outpaced by AI systems.

Weick identified ten reasons why the firefighters couldn't drop their tools. Each one maps with uncomfortable precision onto why professionals today struggle to adapt to AI.

1. Noise that makes listening difficult
   The roar of the fire made it hard to hear the order to change. Today, the noise is different but equally deafening: information overload, competing narratives about what AI can and cannot do, hype cycling next to genuine alarm. Many professionals simply haven't registered the signal clearly yet. The fire is loud, and the instruction to drop your tools is getting lost in it.
    
2. Lack of Justification
   Foreman Dodge gave no explanation for his order. He assumed the danger was obvious. It wasn't. Many organizations are doing the same today, issuing vague mandates to "embrace AI" without explaining concretely what needs to change, why, and by when. People persist in familiar patterns when no clear reason to change has been given.
    
3. Low Trust
   The crew didn't know Dodge well. When an unfamiliar person tells you to do something counterintuitive in a crisis, you hesitate. Many of the loudest voices urging adaptation to AI are consultants, technologists, or executives with whom workers have little relationship. The message arrives without the relational dimension that makes it credible.
    
4. Need for Control
   Keeping a tool, even a heavy one, preserves a sense of agency. For professionals, familiar methods and skills represent a known cause-effect relationship: if I do this, I can produce that. AI disrupts that relationship profoundly. Clinging to established ways of working is partly a rational response to the anxiety of not yet knowing what the new cause-effect relationships are.
    
5. Ability to unlearn fast - Dropping
   One firefighter, running for his life, still stopped to lean a colleague's shovel carefully against a tree. Another spent precious seconds looking for a place to set down his chainsaw where it wouldn't get burned. Dropping is itself a skill that requires practice. Most professionals have never rehearsed what it feels like to set aside a core competency. When the moment comes, the behavior isn't available, because it was never trained.
    
6. Skill with the replacement activity
   Even those who wanted to follow Dodge into the escape fire couldn't make sense of what he was doing. The alternative was more frightening than the familiar danger. For many people today, AI tools are genuinely confusing, inconsistent, and hard to integrate into real workflows. It is not irrational to hesitate when the replacement behavior is opaque and unpracticed.
    
7. Feeling of Failure
   To drop your tools is to admit that what you have been doing is no longer sufficient. That is a profound admission, especially for people whose identity and status are built on years of accumulated expertise. Keeping the tools, continuing to write, analyse, or design the old way, postpones that admission. It lets people feel they are still in the game, still winning, even as the fire closes in.
    
8. Social dynamics
   If the person ahead of you in line keeps their tools, you read that as a signal that things are probably fine. Pluralistic ignorance operates powerfully in organizations: each individual may privately feel anxious about AI, but because no one is visibly panicking or dramatically changing their behavior, everyone concludes there is no need to act. The result is collective inertia built out of individual fear that no one voices.
    
9. Underestimation
   The firefighters didn't believe that dropping their tools would make enough difference. The gain felt trivial relative to the magnitude of the threat. Similarly, many professionals make small, incremental uses of AI, asking it to draft an email here, summarise a document there, without making the deeper changes that would genuinely transform their capabilities. The compounding value of full adaptation is not felt until it's too late.
    
10. Identity
    This is probably the deepest reason, and the one Weick emphasises most. Firefighting tools weren't just equipment. They defined group membership, professional purpose, and personal worth. To drop them was to face the question: without my tools, who am I? The same question haunts professionals today. Writers, lawyers, analysts, designers, programmers. The tool is not separate from the person. It is the person, or so it feels. And that fusion is precisely what makes adaptation so difficult, and so urgent to address.

Advice for People

• Notice which tools are important for your identity. The ones you feel defensive about are exactly the ones to examine most honestly. Defensiveness is a signal, not a verdict.
   
• Practice dropping in low-stakes moments. Try AI on tasks you already know how to do well. Build the muscle of letting go before you are forced to. Address reason #5, skill at dropping, before the crisis demands it.
   
• Distinguish your expertise from your tools. Your judgment, domain knowledge, contextual understanding, and relational intelligence are not replaceable in the same way that execution tasks are. Those are the light tools worth keeping. The heavy ones are the specific methods through which you used to apply that expertise.
   
• Don't wait for certainty. The firefighters who survived didn't have a complete picture — they just moved. Partial adaptation now beats perfect adaptation too late. Counter reason #4 by building new cause-effect relationships through experimentation rather than waiting until the old ones fail completely.
   
• Find trusted peers who are already adapting. Counter reason #3 by seeking out credible voices in your own community, people you know and respect, who are visibly working through the transition. Their example makes the drop feel less like free-fall.
   
• Redefine competence forward, not backward. The question is not "can I still do this without AI?" but "what can I do with AI that I couldn't do before?" This reframe addresses reason #7. It turns adaptation from an admission of failure into an act of professional ambition.

Advice for Leaders

• Give people a reason  to drop their tools. Dodge's fatal mistake was assuming the danger was obvious. It wasn't. Address reason directly: explain the why concretely, connect it to things people care about, and repeat it far more than feels necessary.
   
• Build trust before the crisis hits. Reason #3 is not solvable in the moment. If people don't trust you in ordinary times, they will not follow your unusual orders when transformation feels threatening. Invest in the relationship now.
   
• Make the cost of inaction visible. The 228 feet the South Canyon crew would have gained by dropping their tools was calculable, but no one calculated it for them. Leaders should quantify and communicate what adaptation is actually worth, and what delay is actually costing. Address reason #9 by making the compounding consequences concrete.
   
• Create safe rehearsals for unlearning. Address reason #5 and #6 organizationally. Make experimenting with new ways of working normal, low-risk, and supported. People need to practice the drop and practice the replacement behavior before they need both under pressure.
   
• Disrupt pluralistic ignorance deliberately. Because reason #8 operates silently, leaders need to actively surface the private anxieties people are not voicing. Create forums where uncertainty is named openly, where early adopters share their experiences visibly, and where the social signal shifts from "everything is fine" to "we are all figuring this out together."
   
• Decouple identity from legacy tools institutionally. Reason #1, the deepest one, will not be resolved by communication alone. If your appraisal systems, status hierarchies, and reward structures are still tied to old forms of expertise, you are organizationally incentivizing people to keep their tools. The structural signals must change alongside the cultural ones.
   
• Lead by example. Leaders who are visibly experimenting, visibly uncertain, and visibly adapting give their people permission to do the same. The person at the front of the line sets the signal for everyone behind them.

Conclusion

The deepest lesson from Weick is not about technology — it is about the human tendency to confuse our tools with ourselves. The firefighters did not die because they were cowards or fools. They died because their tools had become indistinguishable from their identity, and no one had ever asked them to practice the difference. AI is simply the latest, and perhaps most powerful, fire to make that confusion potentially fatal. The question is not whether to drop your tools. The question is whether you will do it in time.

Reference

Weick, Karl E.. (1996). Drop Your Tools: An Allegory for Organizational Studies. Administrative Science Quarterly, 41(2), 301–313. doi:10.2307/2393722