The Principle Test

Five questions, run before the excuses begin: could I defend this in daylight in my own name, what compounds if it happens a thousand times, would I accept it from the other side, will I be proud of it in five years, and what does this decision make me. Short enough for one line of a notes app, which is the design constraint.

The pressure it is built for#

When pressure builds, clarity fades. Deadlines shorten, numbers rise, and people start telling themselves stories in which the compromise sounds reasonable. Every professional knows the rhythm: fix it next quarter, raise the concern after the launch, make it right once the board has approved. Drift starts in the background, and by the time anybody names it the reasoning has already adjusted itself to fit.

The antidote is a test that exists before the excuses, and that is fixed enough not to be reworded by the situation. Most ethical failures are not decisions to do wrong. They are decisions taken quickly by people who had no question to put in the way.

The five questions#

1. Public defence. Could I defend this decision in daylight, in my own name? It stops rationalisation, because the reasons that work privately tend to sound different said aloud with your name attached.
2. Scale. If this act were repeated a thousand times, what harm would compound? It reveals the damage that is invisible in a single case, which is where most systemic harm lives and where automated systems operate by default.
3. Reversal. Would I accept this if I were on the receiving end? It restores the perspective of the person affected, and becomes hardest to answer honestly when you are certain they will never find out.
4. Time. Will I be proud of this in five years, when the urgency has faded? It replaces expedience with distance, and urgency is the variable it removes.
5. Identity. What does this decision make me? It moves the question from the act to the person performing it, which is the only version some people will actually answer.

Why five rather than one#

Because each catches a different failure, and any one of them alone can be satisfied by a decision the others would refuse. A decision can be publicly defensible and harmful at scale. It can be acceptable at scale and something you would not accept yourself. The test is cheap enough to run all five, and running one is how it gets performed rather than used.

The scale question and machines#

The second question has changed in force. A decision made once by a person is now often a decision encoded into a system that will make it several million times without anybody revisiting it. The harm that compounds is no longer hypothetical: it is the default behaviour of the thing being built, and the person who wrote the rule may never see a single instance of it. That is the connection between this test and the fourth receipt, which asks who bears the consequence including those who never enter the system as data.

What this has not been shown to do#

The test has not been trialled. Nobody has compared decisions made with it against decisions made without it, and the questions resemble long-standing devices in applied ethics, including publicity tests and reversibility tests, for which no first use is claimed. The claim here is to this set of five, in this order, sized to be run in the moment rather than in a seminar. A person determined to reach a particular answer can pass all five, and the test does not pretend otherwise.

Source#

Rahim Hirji, SuperSkills: The Seven Human Skills for the Age of AI, Kogan Page, 2026, chapter six.

Related SuperSkills research#

• Principled Innovation
• The Integrity Loop
• Rules before tools