← BACK TO CATHOLIC AI GUIDE A CATHOLIC FRAMEWORK FOR ETHICAL AI

The Dignity Filter

There is only one overriding criterion: human dignity.

Every other consideration (efficiency, cost, scale, convenience) is subordinate to this. Technology exists to serve the human person, never the reverse.

THE THREE QUESTIONS

Ask these first. If the answer to any one is wrong, stop. AI should not be used.

01
Does the person end up closer to genuine human connection, or further from it?

AI's highest use is connecting people to real community, real expertise, and real relationship. The test is the net effect on genuine human connection.

02
Would the person on the other end feel deceived?

Transparency is non-negotiable. The question is not merely if the user is aware they are using AI, it's "does the experience create a false sense of being known?"

03
Does this erode human capability?

AI should sharpen human ability, not dull it. If the person using this tool is gradually losing the skill, judgment, or knowledge they once had, the tool is not helping - it is replacing them by degrees.

THE DECISION FLOW
START HERE
STEP 1 Is the bottleneck mechanical? If the real problem is structure, people,
or funding, AI will mask it. Name the root
cause first. Bridge, not bandage.
NO
STOP Not a tech problem
YES
STEP 2 Where does AI sit in the solution? Determine whether AI touches the task,
informs the person, or stands where
a person should be.
RED
STOP AI as substitute
GREEN / YELLOW
GREEN - Tool
YELLOW - Assistant
YES
STEP 3 Are the five obligations met?
NO
FIX FIRST Address gaps, then retry
FIVE OBLIGATIONS
Data SovereigntyTheir data belongs to them. Access, export, delete. No exceptions.
TransparencyNever hide that AI is involved. The person always knows.
ReversibilityThey can walk away without losing their work, data, or capability.
ProportionalityAI only accesses what it needs for the task. Collect the minimum.
Knowing When to StopA tool stays in its lane. Green stays green. Yellow works toward handoff.
YES
PROCEED with confidence
START HERE
↓
STEP 1 Is the bottleneck mechanical? If the real problem is structure, people, or funding, AI will mask it. Name the root cause first. Then decide if AI is a bridge or a bandage.
YES ↓ Continue
NO × Not a tech problem. Stop.
↓
STEP 2 Where does AI sit in the solution? Determine whether AI touches the task, informs the person, or stands where a person should be.
GREEN / YELLOW ↓ Continue
RED × AI as substitute. Stop.
GREEN - Tool YELLOW - Assistant
↓
STEP 3 Are the five obligations met?
YES ↓ Continue
NO Fix first, then retry.
FIVE OBLIGATIONS
Data SovereigntyTheir data belongs to them. Access, export, delete. No exceptions.
TransparencyNever hide that AI is involved. The person always knows.
ReversibilityThey can walk away without losing their work, data, or capability.
ProportionalityAI only accesses what it needs for the task. Collect the minimum.
Knowing When to StopA tool stays in its lane. Green stays green. Yellow works toward handoff.
↓
PROCEED with confidence
Highland • 2026 Based on Catholic social teaching, Antiqua et Nova, and the Rome Call for AI Ethics