Climate and AI: Two transformations, a lot of uncertainty

Two subjects dominate the economic debates of our time: climate, and artificial intelligence.

Both touch your money, your work, your investments. And on both, the numbers are very uncertain. Let's look at them honestly.

First, the cost of climate inaction. In two thousand six, a famous British report, the Stern review, tried to put a number on it.

Doing nothing would cost between five and twenty per cent of world GDP, every year. Acting would cost around one per cent.

Other economists found much lower numbers. Why such a gap? Mostly because of one choice: how much the future is worth.

If you count distant damage as small, inaction looks cheap. If you count it almost as much as today's, it becomes ruinous.

So the range is wide. But almost all estimates agree on one point: damage grows with warming.

A carbon price is the tool most economists judge the most effective. Making people pay for each ton of carbon dioxide emitted.

Say a price of eighty euros a ton. One liter of petrol emits about two point three kilos of carbon dioxide.

In the app, this chapter is animated and narrated, card by card. Start free trial.

That adds about eighteen cents per liter. Nine euros twenty on a fifty-liter tank.

For a factory that emits a million tons, it would be eighty million euros a year. Enough to change its calculations.

Europe has had a carbon market since two thousand five. The price briefly went above one hundred euros a ton, in twenty twenty-three.

The twist: the most effective tool is also the most visible. A price on petrol shows at the pump. Hence its political difficulty.

On to AI. Which jobs, and how fast? In twenty twenty-four, the IMF estimated that about forty per cent of the world's jobs were exposed to it.

And sixty per cent in rich countries. But exposed doesn't mean replaced. Often, some of the tasks change, not the whole job.

On speed, history calls for caution. Electricity took decades to transform factory productivity.

On the markets, AI is already here. A large share of orders goes through algorithms, in a few thousandths of a second.

That lowers execution costs. But machines that react the same way can amplify a panic, as in the flash crash of May twenty ten.

For advice, automatic tools already offer portfolios. Cheaper, but no better at predicting the future.

Finally, concentration. A handful of big tech companies weigh a share rarely seen in American indexes.

A "diversified" index fund can therefore be heavily exposed to a single bet: AI. You can check it in its holdings.

Climate and AI have one thing in common: lots of displayed certainty, few solid numbers. Beware of both extremes.

What you can do: understand the mechanisms, look at what your savings hold, and keep a margin against the uncertain.