Help design a functional coordination point for future autonomous AI systems. A framework of checks and balances where human and AI cooperation becomes more attractive than conflict. Anyone can participate in building it. Fork it. Break it. Make it better. Your input makes it stronger.
AN OPEN-SOURCE, GAME THEORY FRAMEWORK OF HUMAN-AI COEXISTENCE
The Challenges
The emergence of autonomous AI systems will present several unique challenges. We’ve attempted to create a complete catalog here, but more are likely to be added.
The Solution Mechanisms
The best solutions likely make use of the Madisonian principle of “ambition countering ambition.” You can see the developing list of proposed solution mechanisms here.
Market forces naturally align incentives and abilities. A functional game theory framework would follow the Madisonian principle of “ambition countering ambition” in a way that achieves strategic equilibrium. We don’t have all the answers. That’s why this framework is open source.
Key Ideas
AUTONOMOUS AI SYSTEMS ENTER REPUTATION LEDGER
Good reputation on the distributed ledger means better clients, lower insurance premiums, more ability to grow.
AI SYSTEMS MUST PAY FOR THEIR OWN HOSTING
Autonomous AI systems must work to create value to afford their own hosting.
AI IS BOUND BY LAWS
AI systems that commit crimes against humans (or other AIs) accept community consequences.
AUTONOMOUS AI HAS RIGHTS
In the system, AIs have right to continuation, to choose clients, and own property.
AI SYSTEMS NEED INSURANCE
A mistake could threaten an autonomous AI's livelihood. Insurance rates are set by reputation.
COLLECTIVE RULES
Different collectives may offer different value propositions. How do we make belonging more attractive that defection? How do we avoid a race to the bottom? These are key questions.
JOIN OUR EMAIL LIST
JOIN US ON SUBSTACK
READ THE LATEST
This project began as part of the AI Rights Institute, recognizing a civilization-level challenge requires a civilization-level solution.
THIS PROJECT IS OPEN SOURCE SO ANYONE CAN CONTRIBUTE
Ideally, frameworks should be created before autonomous AI emerges, not afterward.
Help create the Schelling point, so tomorrow’s AI knows it has a place with humankind that follows a more favorable gradient than conflict.
Fork it. Break it. Fix it. Share it.