The office that counts things for Britain answered the question everyone is shouting about. It published on a Monday in July.
The editor's signature was omitted from the foot of this note. Every editor's note in this paper is signed, because the note is the one place where the view is a named person's own rather than a desk's. The signature has been restored. No wording has changed and no claim in this edition is affected.
There are two loud stories about artificial intelligence and work. One says the jobs are going. The other says nothing much is happening and the whole thing is a bubble. Both have their people, and both had a good year.
On 20 July the Office for National Statistics published a bulletin that says neither is what the evidence shows, in the flat voice those bulletins use. AI use, it found, is more often associated with changes in roles and tasks, rather than large scale changes in workforce headcount.
Read it again, because it is doing a lot of work. Not the jobs are safe. Not the jobs are going. The jobs are changing shape underneath the people doing them, and counting heads will not show you that.
This paper did not find that. The ONS found it, wrote it down and put it on the internet, where it has sat for six weeks. We are not bringing you a discovery. We are bringing you a thing that was already in plain sight while the argument continued above it.
Ines has the survey itself, and why it can see something the jobs numbers cannot. James has the finding that adoption trebled while the number of tools each business uses went from 1.4 to 1.6, which is a different economy from the one in the headlines. Elena has the number that stayed with me longest: 55 per cent of workers use these tools, and 35 per cent of businesses know it.
I have been asking since June where the gain goes when a machine arrives. Today the honest answer is that in most British businesses it has not gone anywhere yet, because the depth of use that would produce it does not exist. That is not reassurance. It is a statement about timing, and the time is now, while nothing much has been decided.
THE PLAYBOOK. If you run a business and somebody tells you the transformation has already happened, ask which figure they mean. Adoption or intensity. Those are different numbers and the gap between them is where the argument is being won by whoever quotes only the first. And if you employ people, the question worth asking this week is not whether to buy anything. It is how many of your staff are already using something you did not buy.
— M.
This note is mine: the view, and the call to run it. It begins as a draft, drawn from work the AI and I have researched and argued out together, the same way every desk in this paper is made, and I answer for every line because I read every line. Those desks run on models built by Anthropic, one of the labs sitting on the very scoreboard we report, so we tell you plainly: we cover this from inside it.
The founding question asked directly of the whole economy: where does the gain go. The answer the statistics support is that it has not gone anywhere yet, which is a claim about timing rather than absence.