Design an incentive carelessly and people will follow it perfectly – straight off a cliff. Three centuries of history and a widespread current corporate mistake explain why. 

Imagine for a minute that you are a British government employee, stationed in Delhi during the British Raj (that’s somewhere between 1858 and 1947, for those who don’t have their history books handy). You are concerned by the fact that the city is inundated with venomous Indian cobras, but you lack the manpower to tackle that many snakes yourself.

So, you turn to the local populace to help you hunt them down. With the Queen’s blessing, you offer a bounty to be paid for each dead cobra that is presented to you. 

For a while, this appears to be going well. Locals bring you dead snakes, and you pay them. Everyone is happy except, possibly, the snakes, whose numbers are starting to decrease.

But after a few months of this, you notice something strange: it’s the same locals who come forward for payment every time, and even though they are presenting larger quantities of dead cobras, the amount of cobras in the city is starting to increase again.

Eventually, you catch on to the scam: the locals have realised that catching wild snakes is hard work (not to mention dangerous). It’s far easier to breed cobras at home and present them as legitimately wild-caught.

You immediately scrap the bounty, causing uproar among the populace of farmers-turned-cobra-breeders who were enjoying their steady income. With nothing to be earned from their stock, the locals release their broods of cobras into the wild. 

You have spent a small fortune on paying bounties, and you now have more snakes than you had before. Her Majesty will not be pleased. 

There is (unfortunately) no way to prove that this fun little anecdote is a true story. While the foundational narrative may be hard to prove, that didn’t stop economist Horst Siebert from using it as inspiration for a very real phenomenon that he coined “the cobra effect” – which has become shorthand for any situation in which people were unintentionally incentivised to make a problem worse. 

The snakes are everywhere

The trouble with the cobra effect is that once you learn to see it, you start to find it everywhere. As Charlie Munger famously said, “Show me the incentive, and I’ll show you the outcome”. 

In 2002, British officials tasked with suppressing opium production in Afghanistan offered poppy farmers $700 an acre in return for destroying their crops – a staggering sum in a country torn apart by war. Word of the programme spread fast. But the officials had measured the wrong thing. They were paying for destroyed crops, not for a smaller harvest – and so they got exactly what they paid for. 

Farmers planted as many poppies as they possibly could, giving them more crops to destroy and more payments to collect. The craftier ones even managed to harvest and sell the valuable sap before ploughing the plants under, thereby getting paid twice for the same poppies – once by the drug trade, and once by the people trying to stop it. Even the most experienced investment bankers would be impressed by that!

In 2021, the US Congress passed a law requiring sesame – a common allergen – to be clearly labelled on packaged foods, so that allergy sufferers could shop safely. Seems reasonable enough, but the law put food manufacturers in a bind. To sell a product as sesame-free, they now had to guarantee it. In order to be able to guarantee it, they had to scrub shared production lines and continuously test to keep even trace amounts out. That’s a serious expense.

The cheaper option was to give up and go the other way: dump a little sesame into the recipe on purpose, slap it on the ingredients label, and be done. The result was more products containing the allergen, now often added as flour rather than visible seeds, making it invisible to anyone scanning a bun or a biscuit for something they can spot.

A law designed to make food safer made it a lot more dangerous for the people that the law was trying to protect!

The cobra in the org chart

For the past couple of years, executives have been telling their boardrooms a simple story: workers are expensive; AI is cheap.

The strategy? Cut some of the workers, hand the survivors a set of AI tools to make them more efficient, and enjoy the same output at a fraction of the cost. The maths in the financial model (that was probably built using Claude) is irresistible.

Unfortunately, the workers who remain are not greeting their new AI tools with gratitude. Instead, they greet them with suspicion – and reasonably so, having just watched colleagues replaced by the very software they are now being told to embrace. 

A 2026 working paper by Mark Ma and colleagues at the University of Pittsburgh tracked more than 3,200 firms alongside millions of employee reviews. It found that sentiment toward AI turns sharply more negative after a company announces AI-related layoffs, with job-security fears as the single loudest complaint.

That matters a lot, because the same research found that how employees feel about AI is one of the strongest predictors of whether AI actually makes their company more productive. This leads to a company with fewer people and a workforce too wary of AI to get much out of it.

And then comes the bill for the cleanup.

Having discovered that the AI can’t actually carry the load alone, companies are starting to hire the humans back. Staffing firm Robert Half found that nearly a third of companies that cut roles citing AI have already rehired for those same positions. Gartner projects that by 2027, half of the organisations that replaced customer-service staff with AI will do the same.

Add it all up and the ledger is bleak: the company has paid for AI consultants, then for the tech itself, followed by the cost of staff layoffs. Having been through a culture-destroying experience, they’ve then paid to hire them back. The most cynical view is that all these costs end up creating a resentful workforce where the layoffs meant to unlock AI’s value are precisely what buried it.

That is the cobra effect in a modern suit.

You wanted a leaner, faster company. Instead, you taught your own staff to resent the tool that was supposed to save you, crushing years of culture and loyalty along the way.

This doesn’t mean that all AI projects are doomed to fail, of course. It just means that with the wrong incentivisation, you’ll be going from Copilot to cobras faster than you can read the AI-generated restructuring proposal that got you into trouble.

Her Majesty will not be pleased with such a poorly designed reward system. People know exactly how to optimise what you give them. That doesn’t mean that you’ll get the outcome you asked for.

But you will always, always get the one you incentivised.

About the author: Dominique Olivier

Dominique Olivier uses her love of storytelling and ideation to help brands solve problems.

Her first book, Lessons from Loss, has been published by Penguin Random House.

She is a weekly columnist in Ghost Mail and collaborates with The Finance Ghost on Ghost Mail Weekender, a Sunday publication designed to help you be more interesting.

You can learn more about her work at dominiqueolivier.com and she can be reached on LinkedIn here.