Authority at the table · Illustrative research visual
Orientation01 / 06
Why this pathway exists
Human authority must be expressed through risk-tiered permissions, provenance, review, disagreement handling, expiry, and an unmistakable stop.
An AI system can sound considerate while quietly deciding too much. This pathway examines who may propose, approve, override, refuse, and stop—and how the person’s own participation changes over time.
Phase One establishes a research position. It does not claim a deployed care system, clinical authority, or a finished product.
Human oversight
Authored sequence6 papers
The papers in this pathway
Open the memo before leaving the atlas.
The memos are editorial guides derived from the published papers. The originals remain canonical.
These are conceptual studies—not documentary photographs, production screens, or claims of deployment.
01Authority at the tableResearch visual02Human oversightResearch visual03Many hands, one personResearch visual04Review before actionResearch visual
Research remains openPhase Two inquiry
Questions carried forward
What this folder does not pretend to settle.
01
How should supported decision-making be represented when signals or authorities conflict?
02
Which actions can remain open-access, and which require stronger authority?
03
How should the gate mature without freezing an adult inside childhood permissions?
Collaborate with usCareful inquiry
Bring another form of knowledge
Bring perspective to the work.
We welcome careful dialogue concerning Disability rights, supported decision-making, safeguarding, consent, care governance, interface ethics, and accountable public systems.
Please do not send children’s records or sensitive care data in an initial enquiry.
Begin with the research question and the perspective you bring.
