Baseline: the vocabulary and map the family compares itself with
Step 5. The review is finished when it names the baseline.
Shape: baseline
Baseline H1
For vocabulary and map: Akata 2020's hybrid intelligence, with Parasuraman 2000's four functions and levels for who does what. For measuring: Vaccaro 2024's design, which compares human alone, AI alone and the combination.
What the family compares
| Family term | Baseline term |
|---|---|
| person decides | high human control on decision and action selection |
| model proposes, decides nothing | low level of automation on decision selection; the model works at information analysis |
| machine checks | automation of information analysis, with an exact verdict |
| setting the model, machine and person | the collaborative and adaptive challenges of Akata 2020 |
| explaining a verdict | the explainable and responsible challenges of Akata 2020 |
Practice baseline
Agentic AI Foundation 2026 and Anthropic 2024 are the conventions people use now. Both score below the threshold, so they are examples, not support. The maintainer's practice, written down for this review, is the case to compare with them.
How it will be used
- Each member's review adopts the terms above, or says why not.
- Each case reports human alone, model alone and the combination.
- Open: whether the third role needs a name of its own. Parked for the family.