Loophole Ethics

Loophole ethics: why a code of conduct is not a person.

A written policy tells you what an organisation is willing to promise. It does not tell you what it does… the day that promise gets expensive.

Past a certain size, every organisation writes one: an ethics code, an AI policy, a set of principles with a launch event. Publishing it produces a quieter, less advertised effect too. It hands the people inside a specification to satisfy, rather than a compass to steer by.

Once a rule exists in writing, the question has a way of shifting from “is this right” to “does this technically clear the rule as written“. That shift is not new, and it is not unique to AI. It has a name: loophole ethics, the ordinary habit of treating a code of conduct as a puzzle to solve rather than a value to hold!

It looks like a careful reading of the policy, a genuine belief that it was met, and a launch that goes ahead on schedule. The people involved are not usually lying to anyone and that is precisely what makes it hard to catch (with another document.)

The five excuses

When an ethicist raises a finding and gets overruled, the organisation almost never says “we don’t care.”

It reaches for a justification instead, and those justifications fall into a small, recognisable set.

Criminologists studying how ordinary people talk themselves into wrongdoing named these patterns decades ago; they hold up just as well in a product review meeting as anywhere else.

We hear them constantly, in one wording or another.

None of these require anyone to be the ‘bad one’. That is what makes these durable. They let people keep a good opinion of themselves while doing the thing the policy was written to prevent…

Why a policy can’t catch this, and a person can

A code of conduct does not notice when it is being reasoned around, because being reasoned around is exactly the kind of thing a document has no way of seeing -> it sits signed in a governance / HR database and means whatever the room decides it means that day.

That is the argument for the international standard’s ‘Attitudes domain’:

• independence,
• the willingness to escalate, and
• confidentiality that survives pressure.

Not because knowledge and skills do not matter, but because a policy needs a person standing next to it who can hear one of the five excuses forming and say so!

It is also the argument for practising this before the day it matters. Five Objections exists for exactly this reason: each scenario ends not just with what happened, but with a name for the excuse the room reached for when it decided not to listen. Once you can name it in a story, you tend to hear it faster in a meeting.

Want to see what one of these excuses sounds like from the inside? The Minutes takes what was actually said back to you and names which of these four patterns it matches.

Written by us at Ethics Directive. This essay draws on established organisational-ethics research on codes of conduct and on how people rationalise overruling their own better judgement; it summarises that thinking in our own words rather than reproducing any single source. If anything here needs correcting, we will say so in the open, dated.