Free Justice is a public, open, and affordable way to decide together. “Free” here means **free as in freedom**: the rules, tools, and precedents live in the open so anyone can inspect, reuse, and improve them. Published under **legalcommons.org**, Free Justice treats law and policy as a shared craft: continuously updated, versioned, and accountable to the people it serves.
# Why Now We already lowered the cost of finding information to almost nothing. We can now lower the cost of context-aware adjudication to almost nothing, too.
Free Justice uses a Jury of Agents to produce a quick, cheap, *good-enough* provisional judgment inside Living Documents. Human stewards then focus where human care matters most: disagreements, harms, edge cases, and the setting of precedent. Over time, the whole system evolves like a Common Law for the networked age.
# How It Works A case (a policy choice, a dispute, a permit, a procurement) is filed in a living, public document. A jury-of-AIs delivers a **first pass**: a summary of claims and obligations, points of agreement and disagreement, and suggested Deliberative Primitives to apply next (for example, Mirror Role-Swap, Care Ledger Attach, Boundary Proof).
People then deliberate with help from Deliberative Move Recognition—patterns that connect talk to consequences—revise the plan, and record a human ruling with reasons. The ruling becomes a **precedent** that future cases can cite, reuse, or challenge.
# Human-Centred Control Free Justice is designed so that people set the rules and the direction. Curators design transparent training sets. Panels of diverse “judge” models prevent single-source bias. Anyone affected can trigger an appeal.
The arbitration step is not a rubber stamp: it is the place where values are argued openly, obligations are named, and the record is made.
Each accepted ruling updates the shared playbook, much like continuous integration in Policy as Code.
# What “Affordability-First” Means Affordability-first means we do not ask overworked communities or agencies to craft perfection before they can act. We offer a fast, inexpensive **jury pass** that orients the work, then invest human time where it has the highest return: clarifying harms and care obligations, testing boundaries, repairing plans, and writing reasons others can learn from. Instead of fewer, slower decisions, Free Justice enables **many small, well-documented decisions** that improve over time.
# Open by Default Everything that can be public, is public: data provenance, model identities, reasons and dissent, precedent histories, and the metrics we track. Anyone can fork a guide, propose a change, or run their own instance. Free Justice is a commons: free to read, free to use, free to improve.
# Safeguards Free Justice does not hide its fallibility. Each AI opinion is labeled with identity, training snapshot, and confidence. Panels are rotated for diversity. Precedent cannot merge without human sign-off and a short justification. Appeals are easy and expected. Red-teams probe for drift and bias. When in doubt, we pause, fix, and document.
# What Success Looks Like People can see how a decision was made and by whom. Obligations are clear; harms are anticipated and cared for. Plans are revised in light of challenges. Disputes settle faster with better reasons. New cases reuse old wisdom instead of starting from scratch. Over months, the system becomes fairer, clearer, and more responsive—because everyone can help to improve it.
# Start Here Begin with one simple guide in your context: publish a Mirror Role-Swap, a Care Ledger, or a Boundary Proof guide as a Living Documents page. Add a small jury-of-AIs to produce first passes. Keep the arbitration record short, public, and kind. When a case teaches you something new, update the guide. That is Free Justice.
