Human Capability Preservation Standard
AI should strengthen human thinking,not replace it.
AI is becoming capable of tasks that once needed human judgement. Most of the conversation is about what AI can do. HCPS asks a different question.
What happens to human capability when machines can replace it? If AI routinely thinks, decides, and reasons on our behalf, people may slowly contribute less of their own judgement, and lose the habit of it. HCPS exists to define a different path: AI that makes people sharper, not more dependent.
The HCPS principle
Human judgement stays central
The person remains the one who decides. The tool never quietly takes that over.
AI helps surface reasoning
It asks, clarifies, and makes assumptions and uncertainty visible, rather than handing over an answer.
Capability strengthens over time
Used this way, your own thinking grows stronger with use, not weaker, until you have a whole grove of it.
What HCPS looks like in practice
A system aligned with HCPS:
- Asks questions before offering conclusions
- Makes uncertainty visible
- Encourages your own reasoning
- Surfaces assumptions
- Preserves your ownership of the decision
- Helps you think more clearly over time
Arbor Wise: a worked example
Arbor Wise is an early example of HCPS in practice. Instead of generating recommendations, it guides you through a structured thinking process and produces a decision record you control.
Arbor Wise is not HCPS itself. It’s one example of what HCPS-aligned design can look like. It’s being used as a reference implementation while the standard continues to develop.
Why this matters
The direction AI takes will shape how people think, decide, and act. HCPS exists to keep that direction pointed at strengthening human capability rather than quietly replacing it.
HCPS is an open standard
HCPS is still evolving. The goal isn’t to restrict what AI can do. It’s to make sure AI strengthens human capability alongside its own. Read it, challenge the assumptions, and help improve it.
Public draft, version 0.1. Transparency is intentional.