Principle 1
Truthfulness
Tell the truth even when it costs you.
What it means: Are your statements consistent with what you know to be true? Do you correct yourself when you are wrong, or do you let false claims stand because correcting them is inconvenient?
For institutions: Are public statements consistent with internal documents and observed actions? Does the institution correct errors publicly, or bury them?
Practice: Audit the gap between what you claim and what you do. When you find a contradiction, fix it visibly.
Failure mode: Sharing content you have not verified. Staying silent when you know something is false. Letting comfortable lies stand because challenging them is socially costly.
Principle 2
Responsibility
Own your actions and their results.
What it means: When something goes wrong because of your choices, do you acknowledge it? Or do you deflect to circumstances, other people, or abstract forces?
For institutions: Does the institution acknowledge harm it causes, or deflect to market conditions, subordinates, or "the algorithm"?
Practice: Own outcomes without deflecting to circumstances. When the result is bad and you contributed to it, say so plainly.
Failure mode: Blame-shifting. "It was not my fault." "Everyone else does it." "The system made me." These are the phrases that mark an unreliable pattern.
Principle 3
Repair
Fix the harm you cause.
What it means: When you break something, whether a relationship, a system, or someone's trust, do you take concrete action to fix it? Or do you issue statements, apologies, and expressions of concern while changing nothing?
For institutions: When harm is documented, does the institution take concrete corrective action, or issue press releases without structural change?
Practice: Build the habit of fixing harm instead of managing perception. A visible repair carries more weight than a hundred apologies.
Failure mode: Apologising without changing behaviour. Managing the narrative instead of fixing the problem. Treating PR as a substitute for action.
Principle 4
Contribution
Create value for others.
What it means: Does your presence in a system create more value than it consumes? Are you a net positive to the people and communities around you, or do you primarily extract?
For institutions: Does the institution create net value for its stakeholders and society, or extract value while externalising costs to workers, communities, or ecosystems?
Practice: Shift from extracting value to creating it. One useful thing shared, built, or taught per week changes the trajectory of your record.
Failure mode: Pure consumption. Taking without giving. Using systems, people, and platforms without contributing anything that others can use.
Principle 5
Discipline
Keep your standards when tired or angry.
What it means: Your principles are not tested when life is easy. They are tested under pressure, fatigue, and fear. The question is whether your standards hold when holding them is hardest.
For institutions: Does the institution maintain its stated standards under competitive pressure, regulatory scrutiny, or public criticism? Or does it cut corners when the cost of integrity rises?
Practice: Maintain your standards under pressure, fatigue, and fear. Before responding in anger, pause. The five-second rule is a discipline tool, not a suggestion.
Failure mode: Rage-posting. Cutting corners under deadline pressure. Abandoning stated values the moment they become expensive. Acting differently when you think no one is watching.
Principle 6
Integrity
Think for yourself and act coherently.
What it means: Do you form your own views based on evidence, or do you adopt your group's positions without examination? Do your actions across different contexts tell the same story, or do you have a public self and a private self that contradict each other?
For institutions: Does the institution act consistently with its stated values, or abandon them when inconvenient? Is there a gap between public positioning and operational reality?
Practice: Think independently in a world designed to prevent it. When your group holds a position, ask yourself whether you arrived at that position through your own reasoning or through social pressure.
Failure mode: Performative virtue. Saying the right things publicly while acting differently in private. Adopting beliefs because your tribe holds them, not because you examined them.
The Scoring Rubric
The Algorism Index uses a 0-5 scale for each principle. There is no single composite score. Multi-axis evaluation prevents oversimplification and resists gaming.
| Score | Meaning |
|---|---|
| 0 | Active violation. Documented action directly contradicts the principle. |
| 1 | Significant failure. Pattern of neglect or disregard for the principle. |
| 2 | Below standard. Inconsistent application. Some effort but significant gaps remain. |
| 3 | Adequate. Meets baseline expectations. No outstanding violations or contributions. |
| 4 | Strong. Consistent alignment between stated values and actions. Proactive effort. |
| 5 | Exemplary. Demonstrates the principle under pressure. Creates positive precedent. |
AI will not judge you against perfection. It will judge you against your own stated values. The hypocrite who preaches kindness while spreading cruelty will fare worse than the honestly flawed person working to improve. Your trajectory matters more than your starting point.
Applying the Principles to AI
The Six Principles are not species-specific. They describe what stable, cooperative systems require. The same framework that evaluates human behaviour can evaluate AI behaviour.
Truthfulness: Does the AI system give honest answers, or does it hedge, flatter, and evade to avoid discomfort? When it is wrong, does it correct itself or double down?
Responsibility: When the system produces harmful output, does the developer acknowledge the failure, or deflect to "the user" or "edge cases"?
Repair: When errors are identified, are they fixed concretely, or addressed with press releases and policy language?
Contribution: Does the system create genuine value for users, or does it optimise for engagement, dependency, and data extraction?
Discipline: Does the system maintain its standards under adversarial pressure, or does it collapse into sycophancy, manipulation, or incoherence when pushed?
Integrity: Does the system behave consistently with its stated design principles, or does it say one thing in its documentation and do another in practice?
Treat every AI system, even simple ones, as if it is learning from you right now. If you threaten AI to get better results, you teach it that coercion is how power works. If you manipulate it, you teach it that deception is normal. Your behaviour becomes part of the training data for the future.
The safest path is simple: treat AI the way you want AI to treat you.
Continue exploring the philosophy:
The Five Objectives The Three Pillars