TALAMANA · THE AI LITERACY MAP FOR ARCHITECTURE AND DESIGN · Foundations · AGE 15—17 · FACTUAL · DURABLE

People are in the loop

People sit at every joint of the loop, and so does responsibility.

The idea

Human choices run all the way through these systems. People gathered and labelled the data. People tuned the model by rating better and worse answers. People decided where it ships and what it refuses. And a person receives the output. "The machine did it alone" is never the whole story. There are people at every joint of the loop, whether you see them or not.

Why it matters

Wherever there is a person, responsibility can sit. Finding the people in the loop is how you find who answers for what.

See it in the studio

Your rendering tool refuses certain content, prefers certain styles, softens certain edges. Each of those is a person's decision, shipped to millions of users. You are working with strangers through the tool.

Watch for this

The passive voice: "the image was flagged," "the answer was filtered." Someone built the flag. Someone tuned the filter. The grammar hides the people.

Try it

Take one refusal or odd behaviour from a tool you use, and write down which human decision most likely produced it. You will rarely be sure. Notice that too.

Prove it

Trace one output you received through the loop: whose choices touched it before it reached you, and where your own choice enters.

How it works

The tuning step has a name worth knowing: reinforcement learning from human feedback. Thousands of people rate answers, and their preferences are trained into the model's manners. Labelling and rating is real work by real people, and most of it is invisible. What this ring adds to PEOPLE is that the people are specific jobs — labellers, raters, red-teamers, policy writers, evaluators — and each job is a place where a decision can be questioned. The ethics strand of this map returns to them.

What this idea builds on

What this idea opens up

Sources

Open this idea on the map · The complete map · Logika · RBDS AI Lab, India · revised every edition.

Age grows from 11 at the centre to 22 at the edge, and six sectors show the learning strands. Tab into the map and the arrow keys step from idea to idea, following the links where there is one. Enter opens the idea under the cursor, and E reads out its links and the reason recorded on each. Press slash for Search, question mark for the full key list, and Escape to leave. Open Ideas for the complete readable list, including what each idea builds on and what it opens up.

LOGIKA · RBDS AI LAB INDIA
ON-RAMP · AGE 11 · FIRST ENCOUNTERS, NOT GATES — IDEAS · — DEPENDENCIES
DONE
OPENS NEXT
SOLID — STANDS ON