Project Sapier

A CelticFranciscan public-interest initiative

Project Sapier

Human wisdom for artificial intelligence.

AI is getting smarter. The harder question is whether we are becoming wiser. Project Sapier helps business leaders examine AI before efficiency quietly becomes authority, automation becomes dependency, or simulation replaces human relationship.

Take the 15-Minute AI Examination Explore InterfAIth
Project Sapier Human Flourishing Compass Human flourishing at the center, surrounded by six areas of social life affected by AI. HUMAN FLOURISHING THE TEST EPISTEMOLOGY POLITICS ECONOMICS CULTURE ECOLOGY COMMUNICATION
The problem

Most AI decisions are being made too narrowly.

Businesses ask what can be automated, how quickly it can be deployed, and what it will save. Useful questions—but not sufficient ones.

The Sapier question

Does this use of AI increase or diminish human flourishing?

Examine consequences for employees, customers, trust, dignity, agency, accountability, community, and the wider world.

The goal

Examinable AI.

AI should be understandable enough to question, governable enough to constrain, and accountable enough that a responsible human remains answerable.

Before you automate it, can you explain it, defend it, govern it—and live with what it does to people?
Business risk reduction

Move AI from experimentation to governed business use.

Project Sapier helps owners, executives, and boards surface AI risks early—before they become security incidents, customer trust failures, employee disputes, vendor surprises, or expensive remediation projects.

RISKREDUCTIONbefore scale
01

Data, IP & security

Where is sensitive data going? What is retained? What can leak, be manipulated, or be exposed through prompts, integrations, vendors, or agents?

02

Operational & vendor risk

What happens when a model changes, a vendor fails, an agent acts outside expectations, or a critical workflow becomes dependent on opaque automation?

03

People & decision risk

Where does AI influence hiring, performance, pricing, customer treatment, access, reputation, or other consequential outcomes—and where must humans remain accountable?

04

Trust, reputation & oversight

Could leadership explain the system to an employee, customer, board member, auditor, regulator, insurer, or community stakeholder after something goes wrong?

Executive outcome

Know where AI is. Know who owns it. Know what evidence you have.

A Sapier review is designed to produce a practical risk register, accountable owners, recommended controls, and a documented path for escalation—so AI governance becomes part of normal business governance.

↓Unmanaged risk
↑Decision traceability
↑Board confidence
↑Customer trust
NIST AI RMF alignment

A business-friendly on-ramp to Govern, Map, Measure, and Manage.

The NIST AI Risk Management Framework (AI RMF 1.0) organizes AI risk management around four functions: GOVERN, MAP, MEASURE, and MANAGE. Project Sapier uses those same risk-management instincts in plain business language while adding an explicit human-flourishing lens.

NIST's AI RMF is voluntary, not a certification standard, and NIST is currently revising AI RMF 1.0. Its Generative AI Profile adds guidance for risks specific to generative AI. Sapier is intended to help leadership prepare for stronger governance conversations and organize evidence—not to claim NIST certification or legal compliance.

View the NIST AI RMF Run the 15-Minute Examination
AI
RISK
MANAGEMENT
GOVERNOwnership, policy, accountability
MAPContext, stakeholders, impacts
MEASURETesting, evidence, monitoring
MANAGEPrioritize, control, respond
A deeper examination

Six areas of social life AI is already changing

Project Sapier asks leaders to look beyond software functionality and examine the wider system their technology is helping create.

01

Epistemology

What will people believe because of this system? How will we distinguish information, knowledge, and wisdom?

02

Politics

Who gains authority? Who can challenge a decision? Who remains accountable?

03

Economics

Who receives the value created by AI, and who bears displacement, cost, or risk?

04

Culture

What behavior, values, and assumptions will this technology normalize inside the organization?

05

Ecology

What material, energy, infrastructure, and environmental costs sit behind the digital service?

06

Communication

Is AI strengthening human relationship—or becoming a substitute for presence, listening, responsibility, and trust?

Examinable AI

Know it. Question it. Govern it.

1Knowwhere AI is present
→
2Understandpurpose, data & limits
→
3Questionoutputs & consequences
→
4Challengeconsequential outcomes
→
5Ownhuman accountability
Interactive self-assessment

The 15-Minute AI Examination

Choose one meaningful AI use case in your organization. Answer all 15 questions with Yes, Unsure, or No. The questions are designed as an executive screening layer for governance gaps that can feed a deeper NIST-aligned review.

15questions≈ 15 minutes
Question 1
01
PURPOSE
Can we state the specific human or business good this AI use is intended to serve?
“Because AI can do it” is not a sufficient purpose.
Question 2
02
APPROPRIATENESS
Have we asked whether AI is actually the right solution—not merely the newest solution?
Some problems need better process, better leadership, or better human attention rather than more automation.
Question 3
03
PEOPLE
Have we identified everyone materially affected by the system—not only its users?
Employees, applicants, customers, vendors, communities, and people represented in the data may all be stakeholders.
Question 4
04
DATA
Do we know what data the system can access, where it came from, and what may be retained?
Confidential information, personal data, intellectual property, permissions, and provenance matter.
Question 5
05
DECISIONS
Do we know which decisions AI may recommend, influence, or make?
Risk rises sharply when AI touches employment, money, reputation, opportunity, safety, health, or legal rights.
Question 6
06
HUMAN OVERSIGHT
Is there a clearly identified human who can intervene before a consequential outcome occurs?
Human-in-the-loop should mean more than ceremonial approval.
Question 7
07
EXPLAINABILITY
Could we meaningfully explain an important AI-assisted outcome to the person affected by it?
“The model said so” is not a useful explanation.
Question 8
08
CONTESTABILITY
Can an affected person question, correct, or appeal a consequential AI-supported decision?
Accountability without a path to challenge an error is incomplete.
Question 9
09
ACCOUNTABILITY
Can we name the executive or owner ultimately accountable when the system causes harm?
Responsibility cannot disappear into the vendor, the model, or “the algorithm.”
Question 10
10
SECURITY & FAILURE
Have we considered how the system can fail, be manipulated, leak information, or behave outside expectations?
Reliable operation includes graceful failure and an incident response path.
Question 11
11
FAIRNESS
Have we examined whether different groups could experience systematically different outcomes?
Average performance can hide concentrated harm.
Question 12
12
ECONOMIC IMPACT
Do we understand who receives the productivity benefit and who bears the disruption?
Efficiency gains can coexist with loss of skill, dignity, opportunity, or meaningful work.
Question 13
13
RELATIONSHIP
Is AI augmenting human relationship here—or quietly replacing a relationship that matters?
Not every interaction should be optimized into a simulation.
Question 14
14
CULTURE & ECOLOGY
Have we considered what this system normalizes culturally and what material resources it consumes?
Technology shapes organizational values while depending on real infrastructure, energy, water, hardware, and supply chains.
Question 15
15
HUMAN FLOURISHING
Can we make a credible case that this use of AI leaves people and community better—not merely the process faster?
Efficiency is a means. Human flourishing is the end.
Important: Project Sapier is an educational and facilitation framework—not legal, regulatory, cybersecurity, financial, NIST, or compliance certification. The NIST AI RMF is voluntary guidance. High-impact uses of AI may require specialized legal, technical, security, privacy, regulatory, or audit review.
Where this began

InterfAIth

Practices for Hope and Community in an AI-Driven World

Project Sapier grows from the same question at the heart of the InterfAIth anthology: How do we remain fully human while building machines that increasingly imitate us?

InterfAIth gathers stories and practices of hope, community, wisdom, presence, faith, and human flourishing across traditions. Project Sapier carries that conversation into boardrooms, leadership teams, and real-world AI decisions.

Visit the InterfAIth Book Page
01
InterfAIth asks

What keeps us human?

02
The AI age asks

What happens when machines imitate human capacities?

03
Project Sapier asks

How do we make wiser decisions about the technology we have created?

Join the conversation

InterfAIth + Project Sapier Updates

Join the existing CelticFranciscan / InterfAIth community for book launch news, Project Sapier field notes, author conversations, and practical reflections on human flourishing in the age of AI.

For business owners, boards & leadership teams

Want to examine a real AI use case?

Start with one question: Where is AI touching a human being in your organization in a way that could affect their job, money, privacy, reputation, opportunity, safety, dignity, or relationship with you?

Connect Through InterfAIth Follow on Substack
Project Sapier · CelticFranciscan.comTechnology should serve human flourishing, not replace it.