A decision rule for situations where the potential harm is serious or irreversible and the evidence is still incomplete: don't wait for proof before acting with care.
30-second overview
The core idea
Most decisions let you wait for more evidence before acting. The Precautionary Principle covers the exception: cases where the possible downside is serious or impossible to undo, and waiting for full certainty could mean waiting until it's too late. In those specific cases, the principle flips the usual order of operations. Instead of asking "has this been proven harmful?" it asks "has this been shown to be safe enough, given what's at stake?" The burden of proof moves from the people who might be harmed to the people proposing the action.
It is a narrow tool for a narrow situation, not a general license for caution. Used well, it protects against catastrophic, hard-to-reverse mistakes. Used badly, it becomes an excuse to block anything unfamiliar, since almost nothing can ever be proven 100% safe.
Reach for it when
The potential harm is severe, widespread, or affects people who can't consent to the risk.
The damage would be irreversible or very slow and costly to undo.
Scientific or factual evidence is genuinely incomplete, not just inconvenient.
Waiting for more certainty would itself foreclose your ability to act later.
A cheap, proportionate safeguard is available right now.
Skip it when
The decision is easily reversible if it turns out wrong.
The risk is well-understood and already quantifiable (use expected value instead).
The "precaution" itself carries a larger or more certain harm than the thing you're avoiding.
You're using it to avoid ever making a call, rather than to make a careful one.
The stakes are genuinely small, no matter how unfamiliar the situation feels.
Where this model comes from
1970s
German environmental policy adopted Vorsorgeprinzip ("foresight principle"), restricting substances suspected of causing environmental harm, even before the science was conclusive.
1979
Philosopher Hans Jonas, in The Imperative of Responsibility, argued that powerful new technologies demand an ethic that weighs long-term, hard-to-reverse consequences more heavily than short-term benefit.
1992
The Rio Declaration gave the principle its most quoted modern wording: "Where there are threats of serious or irreversible damage, lack of full scientific certainty shall not be used as a reason for postponing cost-effective measures to prevent environmental degradation."
2000s–today
The principle became embedded in EU law covering chemicals, food safety, and GMOs, and is now regularly invoked well beyond environmental policy, in medicine, technology governance, and AI safety debates.
The two moves that define it
Strip away the legal language and the principle makes exactly two moves:
Act before harm is proven, not after. You don't need a confirmed causal link to justify caution, only a credible, serious possibility.
Shift the burden of proof. The party proposing the risky action has to show it's reasonably safe, rather than the potential victims having to prove it's dangerous.
Note what it does not say: it doesn't say "stop all uncertain activity." A serious formulation always weighs the proposed safeguard against its own costs. A principle that blocks everything uncertain blocks itself too, since any precaution is also an uncertain action.
Walk through the checklist
Click each step to see how it works and a short example.
The same model, six different rooms
Worked application: launching an AI feature with an unclear privacy footprint
Common mistakes
When not to use it
The principle earns its keep in a small number of high-stakes, low-certainty, hard-to-reverse situations. Outside that zone, reach for a different tool:
For routine, reversible decisions, a quick pros-and-cons list or a small reversible test beats an elaborate precaution process.
For risks you can actually quantify, expected value or a standard cost-benefit analysis gives a sharper answer than a blanket "better safe than sorry."
For decisions where delay itself causes serious harm, for example withholding a treatment from people who are already suffering, invoking the principle to keep waiting can do more damage than the risk it's meant to prevent.
As a rhetorical trump card to end debate rather than a framework to structure it. If it's being used to avoid ever answering "how much caution, and at what cost," it has stopped being the Precautionary Principle and become a slogan.
Related models in this series
Try it yourself
Describe a real decision you're weighing, and build a short precaution check you can actually use.
Copied
Save this as a reusable skill
This turns the model into a short, portable instruction set you can hand to an AI assistant, or keep for yourself, so you can run the same check on a real decision whenever you need it.
Copied
Apply this skill to my problem
Describe a decision in your own words. This builds a ready-to-paste prompt that applies the Precautionary Principle skill to that exact situation, for you to run with any AI assistant.