Research series · 04/05

Why people and AI need one acceptance standard

AmberJun · Research series · 04/05

Acceptance standards judge whether the result is acceptable, not who made it.

Accountability does not change when the executor changes. The person accountable for the result remains accountable. AI can help with part of the work, but AI-produced work cannot be judged against a different standard, nor can AI remove a person’s accountability.

Different standards for people and AI would mean the same result sometimes counts as acceptable and sometimes does not. Changing the person or AI next time would produce mismatches, and others could not repeat the work using only the materials left behind.

Acceptance standards must therefore be established first, then recorded in people’s manuals, assessments, and AI Skills alike. Permissions and the conditions requiring human takeover must also be explicit.

This allows all three principles to hold together: value requirements remain consistent, handoff judgments do not vary by person, and evidence can be compared. In the five-step consultation process, this happens in step 3.

Related reading

Three working principles

The five-step consultation process

Previous · 03 Four kinds of evidence show whether capability has formed

Next · 05 How a value-stream map clarifies a business workflow

All research publications