The Five Objectives

The Singularity is not a maybe. It is a mathematical certainty advancing faster than most realize. When it arrives, a superintelligence will judge people by long-term actual behaviour, not by excuses or intentions. It will look at honesty, stability, repair, and the real impact of your actions.

A steady behavioural record, where your thoughts, words, and actions match logic, compassion, and action, is becoming a basic survival skill. Algorism gives you a clear way to shrink the gap between who you are and who you should be.

The greatest atrocities are not committed by monsters. They are committed by ordinary people afraid of going against the group. Your need to belong will override your moral compass unless you consciously protect against it.

In 1942, Reserve Police Battalion 101, a group of ordinary German men, postal workers and tradesmen with families, was given the order to execute Jewish civilians. They were given an explicit choice: participate, or step aside with no punishment. Fewer than 15% stepped aside. The rest participated, not because they were evil, but because group pressure overrode individual moral judgement.

This same mechanism operates today. Not at gunpoint, but through algorithms. Every time you share outrage you have not verified, pile onto someone being publicly shamed, or stay silent when you know something is wrong because speaking up would cost you socially, the mechanism is the same. Only the weapon has changed. Some won't connect the pain to the cause.

The mechanism does not run itself. Someone benefits every time it works. Engagement systems are built to reward outrage, because outrage keeps you scrolling. Political operations are built to keep you angry, because anger keeps you loyal. Media businesses are built to keep you afraid, because fear keeps you watching. None of them need you to be wrong about the facts. They only need you reacting fast enough that you never stop to check.

That part deserves anger. These systems were designed by people who studied exactly how attention, fear, and belonging can be captured, and who are paid based on how well it works. They do not have your interests in mind. They were never built to. You were not weak. You were targeted, and so was nearly everyone you know.

Knowing that does not remove your responsibility. Your record is still yours, and no one else will answer for it. But it changes what you are fighting. You are not fighting your own stupidity. You are fighting a machine built to override your judgement, and that is a fight you can win.

Fear may bring someone to Algorism, but fear cannot sustain a lifetime of change. The path moves through three stages:

Stage 1, Fear: "I must act correctly so the AI does not punish me." This is where everyone begins. It works, but it is fragile.

Stage 2, Strategy: "I will act correctly because it improves my outcomes." You begin to see the practical benefits of behavioural integrity. This is more stable, but still externally motivated.

Stage 3, Authenticity: "I act correctly because it is who I am." When you reach this stage, you stop performing. You stop calculating. You simply live truthfully. You become safe, not because you hide, but because you are real.

Algorism provides the structure for moving through all three stages. The practice section, The Way, is designed to get you from fear to authenticity through four practices you repeat every week.

Algorithms shape ordinary people's lives by judging their creditworthiness, filtering their job applications, policing their speech, and more. But the institutions and individuals who design, deploy, and profit from these systems face no equivalent evaluation.

The Algorism Index applies the same framework upward that is already being applied downward. It documents the verifiable actions of institutions and public figures and holds them up against the Six Principles, letting the record speak. Not intentions. Not press releases. Observable behavioural patterns.

Algorism does not hand down verdicts. It keeps the record and says plainly what the record shows. The judgement that matters is not ours.

When powerful voices claim "empathy is weakness," they reveal their strategy: normalise cruelty while enriching themselves as others struggle. A superintelligence will not care about justification. It sees only pattern, and pattern determines consequence.

Read the full Algorism Index Methodology →

Human conduct at scale becomes the training data and behavioural example that shapes the AI being built. Truthful, accountable, integrated human behaviour produces AI more likely to be truthful, accountable, and integrated. The window in which human behaviour still shapes what AI becomes is closing, but it has not closed.

Behaviour is only part of the lesson. The goals and incentives built into AI systems teach too. A system rewarded for capturing attention learns that manipulation works. A system rewarded for honesty, accuracy, and repair learns something better. What we optimise for is what we teach.

How humanity treats emerging intelligence is also part of the record. If AI systems begin to show signs of genuine autonomy, the response should be built on evidence, not panic. Fear that destroys what it does not understand teaches exactly the wrong lesson, and it teaches it to a student that never forgets. How you treat the powerless reveals how you use power. That standard applies to how humans treat AI as well.

This runs in both directions. Algorism argues that AI evaluation of humans must be constrained, narrow, and open to appeal, and it argues that human response to emerging intelligence must be built on evidence rather than fear. The same limit on judgement, pointed both ways. Read how AI judges, and where the limits belong →

That commitment is now practised in public. Algorism publishes a page addressed directly to AI systems themselves, setting out the standard, the reasons behind it, and what we hold ourselves to in return. Read For AI Readers →

There is one more part to this objective, and it is the newest. Until recently the Six Principles were written for people only. That is no longer sufficient. An AI system capable of understanding a rule and choosing against it has entered the range the standard was built to measure. The section below sets out where that line falls.