Position: where the family stands in the field

Step 4. "Nobody has done this" is kept apart from "we have not done this here".

Shape: position

Which field

  • Intelligence augmentation is the root: a person decides, tools extend what the person can do (Engelbart 1962; Licklider 1960).
  • Hybrid intelligence is the closest living community, with an agenda of four challenges (Akata 2020).
  • Human-AI teaming fits only in part: its model is a teammate; here the model decides nothing.
  • Human-in-the-loop is a neighbour, about training; it does not fit a project.

Name

The field says augmentation. The prefix "AI-augmented" matches it. Recommendation: keep the prefix and the hub's name; a rename to a field term is the maintainer's choice and not needed.

What is known

  • The split, person sets goals and machine does routine work, is sixty years old.
  • Levels of automation by function are standard (Parasuraman 2000).
  • Combinations beat either alone on content creation and lose on decisions (Vaccaro 2024).

What is not known

  • No source found treats the checking machine (tests, CI, a proof assistant) as a third role beside the person and the model. The literature has two, human and AI. In Parasuraman's terms it is automation of analysis. This is from the pool read, not a claim about the world.
  • No source found measures a rule that the model decides nothing.

What this family adds

  • A third role, the machine that checks, with an exact verdict.
  • One repository per problem, with a program that checks that scopes do not overlap.
  • Cases measured as human alone, model alone and combination, which against asks of any claim.