Frame: what the field calls a person, a model and a machine on one project
Step 1 of the review, read on 2026-11-02: the problem, the field's names for it, and which one fits this family. It is a first pass of 27 sources, not the full review the sibling topics ran.
Shape: frame
The problem in one sentence
How a person, a language model and a machine that checks share one project, so that the person decides and every claim is checked where it stands.
Terms of the family
| Term | Meaning here |
|---|---|
| person | who owns the project and decides what is claimed |
| model | a language model: it explains, proposes and drafts, and decides nothing |
| machine | the programs that check: tests, hooks, CI, a proof assistant |
What the field calls it
| Name | What the field means | Fits? | Source |
|---|---|---|---|
| intelligence augmentation (IA) | increasing a person's capability to approach a complex problem, with language, artifacts and method | yes: the root, and the family's rule that the person decides | Engelbart 1962; Hassani 2020 |
| man-computer symbiosis | people set goals and criteria; machines do the routine work | yes: the older name for the same split | Licklider 1960 |
| hybrid intelligence (HI) | the combination of human and machine intelligence, augmenting and not replacing | closest: the name of the community that works on this today | Akata 2020; Dellermann 2019 |
| human-AI teaming | people and AI as interdependent teammates toward a shared goal | partly: the family's model decides nothing, so it is not a teammate in that sense | Berretta 2023; O'Neill 2020; Seeber 2020 |
| human-in-the-loop (HITL) | a person inside a machine-learning pipeline: labelling, correcting, steering training | no: it concerns training a model, not running a project | Wu 2022b; Mosqueira-Rey 2022 |
| centaur | a person and a machine playing as one; the popular name | as a metaphor only | Case 2018 |
| levels of automation | which functions a machine does, and how far | yes, for the machine's role | Parasuraman 2000 |
The Engelbart reading
Engelbart names the unit H-LAM/T: a human using language, artifacts and methodology, in which he is trained. Read for this family, language is the model's domain, artifacts are the machine, and methodology is the method of ai-augmented-project. This is the author's mapping, not the source's claim.
Sub-questions
- Q1: what names does the field use, and how many works use each?
- Q2: which name does this family belong to?
- Q3: does a combination of person and model beat either alone?
- Q4: what vocabulary and map can the family adopt rather than invent?
Finding on the repository's name
The field's names are intelligence augmentation and hybrid intelligence. The prefix "AI-augmented" says augmentation, as the field does, and does not say teaming. Recommendation, not applied: keep the prefix; the hub's name stays; a rename is the maintainer's call.