How to use Potentialism Framework.
Use PF to make reasoning more inspectable, disagreement more visible, and revision easier to justify. Do not use PF as proof that a decision is ethical.
PF is a reasoning scaffold, not a clearance badge. It can help structure judgment, but it does not replace evidence, expertise, governance, law, safety practice, or accountable decision-making.
Use PF to improve judgment, not to outsource it.
Use this page when you face a real decision, review, conflict, design choice, deployment, policy, or pilot.
This page turns PF into a practical workflow. It is for moments when you need to reason through what is happening, who is affected, what responsibility looks like, and whether to proceed, pause, narrow, escalate, revise, or refuse.
It can be used before action, during action, or after action — by individuals, teams, organizations, reviewers, educators, facilitators, and people working with AI-enabled systems.
Before action
Clarify the expression, context, affected parties, constraints, foreseeable harms, alternatives, and responsibility.
During action
Watch for new information, pressure, forced urgency, dignity-risk signals, dissent, and escalation points.
After action
Preserve reasons, dissent, outcomes, harms, repair needs, and revision conditions so learning remains possible.
Start with four questions.
Use this when time is limited or when you need a first pass before deciding whether a slower review is needed.
What is actually being done, said, omitted, designed, deployed, rewarded, or enforced?
What power, constraints, incentives, rules, relationships, information gaps, or histories shape the situation?
Who bears risk, who benefits, who has voice, who has exit, and who may be unable to contest the outcome?
What are the likely effects on avoidable harm, dignity of awareness, and agency within real constraints?
Choose the depth of review.
Not every situation needs the same level of process. PF can be used quickly when stakes are low and options remain reversible. It should slow down when stakes, uncertainty, power, dignity risk, disagreement, or irreversibility increase.
Fast mode
Use for low-stakes, reversible, ordinary, or time-limited decisions. Keep it short: expression, context, affected parties, compatibility scan, and a minimal note.
Slow mode
Use when impact is high, disagreement is meaningful, reversibility is low, dignity may be at risk, or power asymmetry is strong. Add documentation, dissent, review, handoff, and revisit conditions.
Temporary action + review
If delay may increase harm but the situation is not clear enough, consider a smaller, reversible, stabilizing move and define when review must happen.
Check what the expression does across three lenses.
Compatibility is not a score. It is a structured way to ask what an expression does to others across time, under real constraints and foreseeable effects.
Avoidable harm
What suffering or harm is foreseeable? What could be reduced by realistic alternatives, smaller scope, better timing, safeguards, monitoring, or repair?
Dignity of awareness
Could the action humiliate, degrade, erase, silence, instrumentalize, or destroy awareness beyond defensible limits?
Agency within constraints
Does the action preserve others’ ability to regulate their potentials, or does it needlessly narrow options, voice, exit, or contestability?
Strength on one lens does not automatically cancel failure on another. “Systemic benefit” is not by itself enough to override dignity or erase avoidable harm.
Use the full workflow when the decision deserves slow-mode reasoning.
This is the more complete version of PF practice. Use it when the situation is high-impact, contested, dignity-sensitive, hard to reverse, institutionally pressured, or likely to matter later.
What is being decided, by whom, and what is inside or outside the current decision boundary?
What is being expressed, through what channel, under what constraints, incentives, roles, histories, and power relations?
Who benefits, who bears risk, who has voice, who lacks exit, and who may be affected indirectly or later?
Which limits are material or safety-based, and which come from incentives, policies, budgets, deadlines, habits, or power?
Can timing, scope, reversibility, safeguards, review, consent, repair, or handoff improve compatibility?
Assess avoidable harm, dignity of awareness, and agency within real constraints.
Where do awareness, power, authority, ability to intervene, answerability, and repair obligations sit?
Keep a trace of reasons, uncertainty, dissent, safeguards, revisit triggers, and repair paths.
Keep enough trace for review and repair.
A PF decision record is useful when a decision is likely to matter later: when stakes are high, precedent may be set, reversibility is low, disagreement is meaningful, or safeguards and monitoring are part of the rationale.
The goal is not bureaucracy. The goal is to make reasoning reconstructable enough that others can inspect, contest, revise, or repair it.
Record what was chosen
Decision, expression-in-context, decision boundary, time horizon, and options considered.
Record why
Compatibility reasoning, constraints, uncertainty, tradeoffs, affected parties, and responsibility map.
Record what could change
Dissent, safeguards, monitoring, revisit triggers, repair conditions, and handoff needs.
Prevent responsibility from disappearing.
In teams, organizations, institutions, and AI-enabled systems, responsibility can drift. Authority may sit in one place, practical power in another, and harm somewhere else.
PF maps responsibility by asking where awareness and power concentrate, who can change course, who will answer for effects, and who may be carrying burdens without real control.
Look for concentration
Where do awareness, power, authority, technical control, institutional leverage, or information access concentrate?
Look for mismatch
Where do formal authority and practical power differ? Who is expected to answer without being able to change outcomes?
Look for drift
Where might delegation, automation, hierarchy, metrics, or process make responsibility evaporate?
Look for repair
Who can pause, narrow, reroute, refuse, escalate, compensate, repair, or revise the system after effects become visible?
Use PF across the life of a decision.
Before action
Name the expression, context, affected parties, constraints, options, likely harms, dignity risks, agency effects, and responsibility holders.
During action
Watch for changing facts, forced urgency, new affected parties, dissent, escalation signals, irreversibility, and dignity threshold risks.
After action
Compare expectations with outcomes, record harms and blind spots, preserve dissent, trigger repair, and revise future practice.
Hand off when the situation needs more than ethical orientation.
PF is insufficient by itself when a situation requires legal authority, clinical expertise, technical safety assessment, security review, formal governance, domain evidence, emergency response, or affected-party process with real power.
PF can help name why a handoff is needed. It should not pretend to be the handoff.
High stakes
Slow down, document more carefully, widen review, and involve relevant expertise or authority.
Technical risk
Pair PF with safety engineering, security review, technical assurance, testing, monitoring, and failure analysis.
Legal or clinical stakes
Use qualified professional, legal, clinical, institutional, or emergency channels where appropriate.
Watch for PF becoming theater.
Checklist theater
“We filled the form, therefore the decision is ethical.” PF artifacts are supports for reasoning, not proof.
Compatibility-washing
Using PF language to justify a preferred decision while hiding harm, dignity risk, dissent, or responsibility gaps.
Forced urgency
Treating speed as unavoidable when urgency may be produced by incentives, authority, prior delay, or avoidable design.
Start with one real decision.
Choose a decision, conflict, policy, deployment, or pilot. Do the one-minute scan first. If the situation is high-stakes, contested, irreversible, or dignity-sensitive, move into slow mode and create a decision record.