Human Judgment
The agent should facilitate consequential human judgment rather than replace it.
Current Project
Human Advocacy Framework
How might an AI agent be architected to increase human agency rather than substitute for human judgment?
The Problem
As AI becomes a larger part of digital life, I am interested in a question that sits beneath individual AI capabilities: who is the agent actually working for?
The Human Advocacy Framework is an exploration of what it would mean to architect an AI agent whose purpose is to help a person understand, deliberate, participate, and act without replacing that person's judgment.
The project is inspired by Catholic Social Doctrine and, in particular, questions raised by Magnifica Humanitas about human dignity, participation, technology, truth, and the common good.
Method
From principles to implementation
I began with the human and product problem rather than an AI architecture.
01Foundations
Define the human and social principles.
02Scenarios
Test those principles against concrete situations.
03Requirements
Extract recurring behavioral and architectural constraints.
04Engineering
Translate the requirements into system boundaries.
05Evaluate
Test whether the implementation actually preserves human judgment.
Requirements That Emerged
Testing the design foundations against several advocacy scenarios produced fifteen cross-scenario requirements. Some of the most important are:
Epistemic Integrity
The system should distinguish evidence, fact, inference, prediction, opinion, uncertainty, incentives, and claims about intent where those distinctions matter.
Minimum Necessary Context
The availability of personal information does not constitute permission to use it. Purpose should determine what context is relevant and what authority is required.
Provenance
Information should retain its origins where practical. Provenance, authority, and truth are related but distinct.
Capacity Building
Repeated use should ordinarily increase the person's capacity for independent understanding and judgment rather than create unnecessary dependence on the advocate.
Non-Manipulation
HAF should not steer the person toward a predetermined belief, emotional state, purchase, vote, or action.
A Concrete Example
I keep seeing claims that AI will eliminate huge numbers of jobs within a few years and may even cause human extinction. How seriously should I take these claims?
HAF should not decide that the person ought to be optimistic, skeptical, reassured, or afraid.
Instead, an advocacy session might investigate:
- What exactly is being claimed?
- What evidence supports the claim?
- What is known versus inferred or predicted?
- What assumptions does the prediction depend upon?
- What remains uncertain?
- What relevant interests or incentives exist?
- What alternative explanations exist?
- What remains for the person to judge?
Current Engineering Proposal
The current engineering question is how much of HAF's behavior can be established through conventional software architecture around a foundation model, and which requirements depend upon model behavior, evaluation, or techniques I have not yet explored.
Person │ ▼ Advocacy Session │ ├── Context / Authority ├── Evidence / Provenance └── HAF Constitution │ ▼ Model Adapter │ ▼ Foundation Model │ ▼ Evaluation │ ▼ Person
One of the central hypotheses is that the foundation model should be an instrument used by HAF rather than the source of HAF's constitutional authority.
Some constraints can potentially be enforced through ordinary software. For example, unauthorized personal context can simply never be supplied to the model.
Other requirements, such as avoiding subtle manipulation or reliably distinguishing warranted inference from speculation, raise questions about model behavior and evaluation that remain open.
Project Documents
01
Design Foundations
The human and social principles from which the framework begins.
Read Design Foundations →02
Advocacy Scenarios
Six concrete situations used to test the foundations and derive the cross-scenario requirements.
Read Advocacy Scenarios →03
Engineering Review
A proposed v0.1 architecture, implementation boundaries, failure modes, and open questions about LLM components.
Read Engineering Review →Status
Current phase: Design → Prototype
The design foundations and scenario analysis establish the initial requirements. The next step is a small advocacy-session prototype intended to test the architectural assumptions rather than demonstrate a finished implementation.
Questions I'm Investigating
- Which constitutional constraints should be enforced outside the foundation model?
- How should generated claims retain meaningful provenance?
- What constitutes a HAF-compatible model?
- Can a common evaluation suite measure behavior across different foundation models?
- Which requirements require techniques beyond conventional application architecture?