PAICE Principles: What PAICE stands for, and what it refuses to become

PAICE.work PBC exists for one reason: to make AI collaboration measurable, teachable, and governable.

Not in theory. Not in keynote slides. In actual workflows, with actual risk, and actual consequences.

These principles are how we keep PAICE useful in the real world, even as tools change, laws change, and cultural narratives drift.

Principle 1: People first, because impact lands on people

PAICE is designed around human stakeholders. People are the ones who carry responsibility, experience harm, make decisions, and live with the outcomes.

When PAICE says "people," it means humans. We're primarily focused on those stakeholders inside an organization's duty-of-care and governance boundaries, but we mean all people.

Principle 2: AI is non-human, and must be governed as a system

PAICE treats AI as non-human computational systems used to generate, recommend, decide, or act. That includes models, agents, and AI-enabled workflows.

PAICE does not romanticize AI. It does not anthropomorphize it. It does not outsource accountability to it.

AI can be powerful without being treated as a person. In PAICE, AI is governed as a system. And People+AI is also treated as another larger system.

Principle 3: Accountability is non-transferable

If an organization uses AI, the organization remains accountable. PAICE does not offer a way to shift blame onto tools, vendors, or "the model."

If you deploy AI into a workflow, you still own the outcomes. PAICE helps you prove you can own them responsibly.

Principle 4: Measurement beats mythology

AI conversations are crowded with vibes, hype, fear, and creative storytelling. PAICE lives on the opposite end of that spectrum.

PAICE measures what is actually happening:

Where AI shows up in workflows
How it is being used
Where value is being created
Where risk is accumulating
Where governance is missing, unclear, or performative

We do not need perfect certainty to make progress. We need observable reality and measurable improvement.

Principle 5: Governance is a capability, not a document

Policies are not governance. Training is not governance. A committee is not governance.

Governance is a capability. It is the organization's ability to repeatedly produce safe, aligned outcomes under ever-changing conditions.

PAICE assesses that capability, identifies gaps, and makes improvement measurable.

Principle 6: Safety is designed into workflows, not added as a warning label

Most "safe AI" efforts stall at guidance and good intentions. Meanwhile, real work routes around them.

PAICE prioritizes workflow design because workflows are where behavior becomes default. If a normal person can complete the workflow without knowing where the AI step lives, AI is not embedded. It is merely a demo.

PAICE helps teams move from heroic individual usage to stable, governable defaults.

Principle 7: Privacy and data minimization are features, not footnotes

If a measurement system requires excessive data capture, it becomes a liability instead of an asset. PAICE is built to surface patterns without turning organizations into surveillance machines.

Collect what is necessary. Protect what is collected. Retain only what you can ethically & legally defend retaining.

Principle 8: PAICE is policy-independent, not values-neutral

PAICE is a measurement and governance framework. It does not advocate for or against legal personhood for AI.

It is compatible with different legal regimes because the core question stays the same: Can your organization collaborate with AI systems safely, responsibly, and measurably, under applicable law?

PAICE has values. It is not neutral about accountability, duty of care, or human impact. But it does not exist to win philosophical arguments. It exists to help organizations run AI in a mature and defensible way.

Principle 9: Capability should be improvable, not punitive

PAICE is not a gotcha. It is not a compliance trap. It is not a purity test.

The goal is improvement and workforce enhancement, not shame or workforce reduction.

A useful assessment creates clarity, reduces friction, and supports better decisions. PAICE is designed to do that, repeatedly.

Principle 10: The output must be actionable

A score with no next step is entertainment. PAICE is not entertainment.

Every PAICE result should produce:

Clarity about current reality
A shared vocabulary for leaders and teams
Prioritized gaps worth fixing
A path to measurable improvement
Board-usable reporting language when needed

If the output does not change behavior, it is not finished.

What this means in practice

PAICE will always prioritize:

Measurable workflows over vague intentions
Accountability over automation theater
Governable defaults over heroic effort
Human impact over novelty
Clarity over ideology

That is the work. That is the point.

If you want the short version, it's this: PAICE measures whether your organization can collaborate with AI systems without lying to itself about risk.

PAICE Principles