Non-negotiable
01
Conditions, not merit
A ledger may audit the circumstances under which work gets made. It may never grade the work itself.
DRAFT SPEC · v0.2
The Human Ledger
A disclosure standard for the conditions under which human work gets made in any organisation deploying AI. It never passes a verdict on the work itself.
What this is
If we measure AI success solely by head-count reduction and output volume, we fall victim to the exact trap Kennedy warned us about in 1968. But if we design our implementations to promote human agency, we turn technology into a tool that serves the human experience rather than exploiting it.
The Human Ledger is an attempt at the instrument that would make that difference visible. It has three layers, borrowed from the way the GHG Protocol is built:
- Five principles, which make the argument.
- Six core conditions, which are what an organisation actually discloses and are identical in every sector. This is the standard.
- Sector annexes, which work the core through one sector’s P&L. Creative, marketing and media is written, and five more are open.
Systemic bug
We identify an easy-to-measure metric → we bend our entire culture to maximise it → we destroy the qualitative things we actually care about.
That bug is why the principles come first. Somebody proposes a metric, everybody argues about the metric, and the argument slowly settles the principle without anyone having agreed to it. So the principles come first and the numbers second, and the numbers are negotiable.
The canon
Five principles
02
Publish everything you measure
The opposite of a human ledger is not an unmeasured world. It is a world measured privately.
03
Floors, not targets
There is no upside to exceeding a floor, so there is no incentive to fake one.
04
Count what you export
Offshored volume has to land on the ledger of whoever commissioned it, or the ledger becomes a laundering mechanism with a nice logo.
05
Protect the unmeasured space
Some work only proves its worth later. Disclose how much room people have for it, and never measure what they do inside it.
↗
All five, expanded
With the full argument for Principle One, and the two metrics it forced us to withdraw.
The standard
Six conditions
Hours, training, junior headcount, human authorship, rework and exposure: externally observable, the same in every industry, and set as floors.
Unfinished
Four arguments I can’t finish on my own.
These questions are open, and not as a figure of speech: I haven’t resolved any of them. If you have the answer, it belongs in the debate rather than in my drafts folder.
Question 01
Who holds the pen?
There are only three answers. A government, and you get direction. Capital, and you get liability management. Nobody, and you get exploitation with good quarterly results. I don’t think there’s a fourth. My own view is that the pen belongs to a legislature for the disclosure standard and to nobody at all for the quality judgement. But that’s a position, not a settled fact.
Argue this on LinkedIn ↗Question 02
Which of the six conditions survive?
I suspect some of them collapse in the debate. I’d like to find out which, and I’d rather hear it from the people who would have to file the numbers, in a hospital, a law firm or a bank, than from people who simply like the idea of them.
Argue this on LinkedIn ↗Question 03
Who audits it?
D&I and sustainability reporting only grew teeth when disclosure became mandatory and third-party assured. Self-reported human metrics are marketing.
Argue this on LinkedIn ↗Question 04 · the hardest
What stays permanently out of scope, and who guards the boundary?
This is the one I’d most like help with, because a boundary that can be moved by a single majority isn’t a boundary.
Argue this on LinkedIn ↗What’s changed
v0.1 → v0.2 · 13 October 2026
This is a working spec, so the thinking moves, and when it does the old wording stays visible, struck through, with the date and the reason. Here is the latest change, with the earlier ones a click away.
Part I restructured around the six conditions
Was
Part I argued from the creative industries: Act III was written from the marketing front line, and Act IV ended with the ten-row creative table as the worked example.Now Part I argues from work in general and closes Act IV on the six core conditions. The creative-industry material, including the original Act III and the table, sits unchanged in an annex at the end.
Earlier changes (9)
Principle Five clarified: measure the size of the space, never what happens inside it
Was
Premature, difficult and illegible work needs somewhere to survive until it becomes recognisable, and that describes most of the work that ever mattered.Now Some work only proves its worth later. Disclose how much room people have for it, and never measure what they do inside it.
Exposure defined
Was
A one-line definition that measured operator time, and so duplicated Rework.Now Exposure: the AI-mediated adverse events that land on people, counted by class of harm and disclosed by position relative to the system.
Judgement rate absorbed into Exposure
Was
Judgement rate as a freestanding metric: a row of the Annex A table and Part II’s central number.Now Judgement rate is the P2 cell of Exposure: the operator position.
Principle Three clarified: floors govern process, trends govern outcomes
Was
Floors, not targets, with nothing said about which measures can carry a floor.Now Procedural cells take real floors. Outcome counts are disclosed with their trend and never targeted.
The six conditions are a set, not a menu
Was
Six conditions, with nothing said about reading one without the others.Now Rework is read alongside human authorship and exposure for the same period. The core is a set, not a menu.
Trust delta is withdrawn
Was
Trust delta: what happens to “trustworthy” and “authentic” as our AI content share rises.Now Withdrawn. It breaks Principle One.
Originality distance is withdrawn
Was
Originality distance: distance from category convention, and from our own last 12 months.Now Withdrawn. It breaks Principle One.
Six conditions become the core of the standard
Was
Five principles, and ten metrics offered only as illustration.Now Five principles and six disclosable conditions: hours, training, junior headcount, human authorship, rework, exposure.
The standard is cross-industry
Was
A disclosure standard for the industries that manufacture attention, measuring the conditions of creative work.Now A disclosure standard for the conditions of human work in any organisation deploying AI. One universal core, with sector annexes on top.
Sector annexes
One written. Five open.
An annex is a worked example: the same six conditions applied to one industry’s real P&L. The core stays the same from sector to sector and only the annex changes. The open ones have been left as gaps on purpose, because they need people who work in those sectors to write them.
Annex A · Written
Creative, marketing and media
Agencies, brands, studios and publishers: the canary rather than the subject.
Open
Health
Hospital trusts, clinics, primary care, diagnostics.
Open
Legal
Law firms, in-house legal teams, courts and tribunals.
Open
Finance
Banks, insurers, asset managers, accountancy.
Open
Software
Product and engineering teams, IT services.
Open
Education
Schools, universities, training providers.
The list
The ledger is a working spec, not a manifesto.
Get new principles, revisions and the arguments that killed them, roughly once a month and never more often.