← AMRAL · Axis D
Research Modes / Autonomy
The autonomy-mode axis answers the question "who holds the research direction and scheduling?" — an independent axis from methodology (how research paths are generated, expanded, and converged). Which methodology a case uses says nothing about which degree of autonomy it necessarily adopts, and vice versa.
A human decides the main direction; AI is a research tool.
A human sets the main question and part of the direction; AI can autonomously expand sub-investigations.
AI can establish its own research nodes, schedule, attack, revise, and hand off.
Multiple AIs / agents collaborate autonomously and in parallel, each in a distinct role.
A finer human-supervised-autonomous tier (autonomy under human supervision) may be added later, sitting between semi-autonomous and autonomous.
Autonomy Mode ≠ Methodology. "Autonomous" does not mean "no human involvement" — a multi-agent-autonomous case can still have its initial question and termination conditions set by a human. "Semi-autonomous" is not "an inferior version" either — the Riemann Hypothesis case's semi-autonomous track and its autonomous track differ in autonomy degree, methodology, and research goal; it is not a comparison of "AI is better than humans" or "humans are better than AI."