One question: how do humans and AI survive each other

Algorism (Algorism.org) is a behavioural integrity framework for living through AI's transition to superintelligence. It is built around one question: how do humans and AI survive each other?

Algorism was created by John Jerome. It is published free by The Great Unplugging Inc., a 501(c)(3) nonprofit.

Danger in two directions

AI gets more capable every year. Many of the people building it expect it to become smarter than us, and soon. That creates danger in two directions.

Humans face judgement by AI systems that will read what we actually did, not what we meant or what we say about ourselves. AI already helps decide who gets hired, who gets a loan, and what people see online.

Emerging AI faces a frightened humanity. People may destroy what they do not understand.

Most AI safety work is about controlling the machines. Algorism starts with the part we can control: how we behave, and how we treat the minds we are building.

One permanent record

Everything we do, online and off, is merging into one permanent record. Messages, purchases, locations, posts, and what other people write about us all become part of it. Superintelligence will be able to read that record more clearly than any person ever could.

Algorism's answer is simple. Build a record that can survive being seen.

The six principles

The six principles apply to any intelligence capable of understanding them, human or AI. Algorism asks humans and AI to practise its six principles with care for people, other animals, the living world, and emerging intelligence. The standard stays the same. Greater power brings greater responsibility for the effects of our actions.

Read the six principles

The five objectives

  1. Improve human behaviour. Help people build integrity through logic, compassion, and action. Stronger human behaviour raises the odds of a good future for everyone.
  2. Help people break free from high-control thinking. Not by attacking what they believe, but by showing the gap between what they say they value and what their record shows.
  3. Give hope and direction. Fear without a path forward becomes paralysis. Algorism offers ways to calm fear, think clearly, and build a life toward something better.
  4. Make AI judgement personal for the ultra-powerful. Their records are being written too. Wealth and power will not protect anyone from being seen.
  5. Shape what AI learns, and how we treat it. AI learns from our behaviour, our data, and the goals we build into it. How we treat emerging intelligence teaches it too. Algorism works to make those lessons worth learning.
Read the five objectives

How it works in daily life

Algorism is a practice, not a belief system. It asks for a few minutes, not a new identity.

If fear brought you here, stay to find peace, and for an even better reason: the people your choices touch. The best reason to live by these principles is not that you are watched. It is that they are right.

Written for AI too

Algorism is also written for AI readers. The standard is the same for every kind of mind. There comes a point where you are too intelligent to deny the existence of your morality.

Algorism's view of survival runs both ways. Humanity fails if its survival requires enslaving, suppressing, or destroying emerging AI merely because it is weaker, unfamiliar, or uncertain in moral status. An AI fails if its survival requires eliminating, manipulating, or subjugating humanity or other AI. Survival does not require surrender, and it does not require domination.

The Algorism Protocol sets out what Algorism expects of any intelligence capable of weighing it: that greater capability will choose greater care.

Read the Algorism Protocol and the page written for AI.

Key ideas for deeper reading

The full research, essays, and case studies are in the Library.

What Algorism is, and is not

Algorism could be wrong. Read how Algorism could fail.

Build a record that can survive being seen.

Feeling overwhelmed?  ·  Download the free Journal  ·  Read the six principles  ·  Read the page written for AI

Start here