This week the biggest companies in the world say how much they are spending on AI.

On Wednesday and Thursday, four of the world's biggest companies publish their results. The Federal Reserve sets American interest rates on Wednesday. The Bank of England sets ours on Thursday.

Every one of those rooms will ask a version of the same question. You are spending an enormous amount of money on artificial intelligence. When do we get it back?

Give the analysts this much: asking is their job. They are paid to work out whether a company is worth owning, and a company that has doubled its capital spending in a year and spent more cash than it made is exactly what they should be pointing at. Alphabet did that on 22 July and the market marked it down anyway, on a quarter where nearly every other number came in ahead. That is a market doing its job.

But notice what the question quietly assumes: that the only person with a stake here is a shareholder. These are machines bought to do work that somebody currently gets paid for. Nobody in those rooms on Wednesday or Thursday will ask whose.

British American Tobacco stood up on 29 June and said it was changing about nine thousand roles. Every headline the next morning called it an AI restructuring. Read the company's own release and the words artificial intelligence do not appear in it. Not once. What it says is technology enabled. What it describes is thousands of jobs moving to Accenture and two other outsourcing firms.

So we do not know. Maybe a machine is doing that work now. Maybe somebody in Costa Rica is. The company has not said, it has not said how many of the roles are British, and there is no number anywhere that would settle it. It publishes results on Thursday. Somebody should ask.

Nobody on our front bench will. Nor the one opposite. Roll it back through the last three governments and the shape does not change: a great deal about being a leader in this, almost nothing about being a country where people work. We have a Minister for AI sitting at the Cabinet table now, a first. The department that had been building the machinery was abolished in the same week, and the AI brief was cut into pieces and handed to two others.

That is not one party's failure. That is a whole political class treating this as an industrial policy question and not a question about people's lives.

We exist to ask one question. What do humans do when we don't have to work? This week the world spends four days totting up the bill for not having to, and it will not once occur to anybody in the room to ask the second half.

THE PLAYBOOK

Ask what your employer's AI budget is meant to replace. Not whether they are using it, everyone is. What the spending is for. The answer, or the absence of one, tells you more than any company announcement will.

Read the headcount line, not the profit line. Listed companies publish the number of people they employ alongside the money. It sits in every results statement and it is the number nobody reads out.

If a data centre is planned near you, look up its grid date. Announced and connected are years apart in Britain right now, so local promises of jobs are dated by the connection queue, not by the press release.

The Playbook is general information. It is not financial, legal or career advice.

— M.

This note is mine: the view, the argument, the call to run it. It begins as a draft, drawn from a body of work the AI and I have researched, debated and explored together, the same way every desk and article in this paper is made. Nothing publishes until it says what I actually mean, and I answer for every line, because I read every line: I push back where I disagree and agree where I don't. Those analysts all run on models built by Anthropic, whose industry we cover, so we tell you plainly, we report on this from inside it. The Quernal is one human working with AI to do what once took a newsroom, held to the standard by the gates we built: nothing publishes until it has passed them, and I, Matt Brazil, am that gate.