Skip to content
Part of Human Value in AI ↗Family overview
Human Creative Value in AI

Evidence papers

Read the papers

Cited papers for HR leaders and internal proposals. Claims are labelled as published evidence, our interpretation or an unvalidated proposed method.

Free · full paper · 5 pages

Before the Prompt

Judgment, framing and calibration as human contributions to AI-assisted work

No sign-up, no email. Not peer reviewed.

A person can see an image is almost right without knowing Photoshop. This paper argues that seeing what fits and making it are separate skills. It does not claim judgment is innate or AI review is useless.

  • Fact Six published sources, each with a working DOI link.
  • Illustration The undercooked steak returned burnt: a direction is not a target.
  • Interpretation Overcorrection, regression, drift and repetition need different responses.
  • Planned A calibration audition for hiring. Proposed, not validated.

In preparation

  1. 2

    Calibrate, Don’t Overcorrect

    Tell overcorrection, regression, drift and repetition apart. Write a revision brief that keeps what is right.

    $15
    Not yet on sale

  2. 3

    The Cost of Another Revision

    What repeated corrections cost, when to stop and how to make an internal case with clear evidence labels.

    $15
    Not yet on sale

These papers are still being written. Nothing can be bought here yet. Once secure payment is set up, each will sell separately. No subscription.