Evaluative agency: what the person must keep

AI may help produce the work while you keep responsibility for judging it. Evaluative agency names that responsibility.

About 20–30 minutes · 13 named steps · research detail is optional

The previous step showed why immediate performance and later capability can move in different directions. Evaluative Agency Theory explains the difference by separating production from evaluation.

Producing an artifact and deciding whether it is adequate are different acts. AI makes production dramatically cheaper. It can help with evaluation too, but delegating both removes the comparison, checking, and revision through which people exercise judgment.

Evaluative Agency Theory adds an AI-specific claim in three parts.

First, separability. Producing an artifact and evaluating it are different acts, and AI changes their relative cost by different amounts. Production becomes dramatically cheaper. Evaluation becomes only somewhat cheaper, and can become harder, because the person is judging work they did not fully construct. Research on cognitive offloading and levels of automation gives that separation a broader foundation.

Second, locus. Evaluation is the act that builds and expresses judgment. Delegating production can preserve that act; delegating evaluation removes the repetitions that would have developed it. Third, cultivability. Evaluative agency can grow through repeated, reflective practice and weaken when it is continually displaced. No single prompt proves either change.

Put together: the useful question is not how much work was delegated. It is whether the person retained authorship of the evaluation while using AI to change how the work was produced.

Why this predicts the split in the literature

A theory earns its keep by explaining results that already exist without being retrofitted to them.

Evaluative agency predicts that immediate performance and later capability come apart under AI assistance, in that specific direction: output up, retention flat or down. It predicts that the gap widens as the delay to the unaided test grows, because what is being measured at three days is closer to the artifact and what is measured at three weeks is closer to the person. It predicts that an intervention which forces evaluation — reflection, verification, having to defend a choice — narrows the gap without giving back the production time.

Those are the shapes the studies on the previous step actually have. Which is evidence for the theory, not proof of it: several other accounts predict the same shapes, and separating them is what the research program is for.

The shape the theory predicts, drawn as a shape. The artifact stays good; the person diverges from it, and the divergence grows with distance from the task. Keeping the evaluation is what moves the third line. Illustrative. The axes carry no units and none of these curves is a dataset — the studies on the previous step have this form, not these values.

What evaluative agency rests on

These sources establish the separation between production, monitoring, control, testimony, automation, and dialogue. They do not by themselves validate the complete AIEM pathway.

Supports Mechanism

The extended mind
What it established. An external resource that is reliably available, easily accessed, and automatically trusted can be part of a cognitive process rather than merely an input to it.
What it licenses here. The mechanism layer's insistence that the unit of analysis is the person-and-tool system. It does not license the stronger metaphysical reading of the thesis, which remains contested; the model uses the three criteria as a description of when coupling is tight, and the third one — automatic trust — is exactly what evaluative agency withholds.
Clark, A., & Chalmers, D. (1998). The extended mind. Analysis, 58(1), 7–19. Source
Mixed-initiative interaction
What it established. The useful designs are not full automation but a negotiated division of labor, in which either party can take the initiative and the system reasons about the cost of acting versus waiting.
What it licenses here. Reading an interaction as co-produced, and reading the modes as answers to who took the initiative for what. It is a design principle for interfaces, not an empirical finding about learning, and nothing about outcomes follows from it.
Horvitz, E. (1999). Principles of mixed-initiative user interfaces. Proceedings of CHI '99, 159–166. Source

Supports Dispositions

Epistemic dependence
What it established. Much of what any person rationally believes rests on the testimony of others whose reasons they cannot personally check. Dependence is the normal condition of knowing, not a lapse in rigor.
What it licenses here. The model's refusal to treat reliance as a deficit, and its choice of calibration rather than independence as the disposition that matters. Hardwig is arguing about human experts, who can be held responsible in ways a model cannot; the model borrows the structure of the argument, not a verdict about AI.
Hardwig, J. (1985). Epistemic dependence. The Journal of Philosophy, 82(7), 335. Source

Supports Principles

Metacognitive monitoring and control
What it established. Cognition has two levels. An object level does the work, and a meta level monitors it and issues control decisions back down. Monitoring and control are separate functions and can fail separately.
What it licenses here. The claim that engagement is controllable rather than merely observable, and the shape of a principle as a standing control policy. The separation also explains a specific failure: a person can monitor an AI answer accurately and still not act on it, and the model has to treat that as different from not noticing.
Nelson, T. O., & Narens, L. (1990). Metamemory: A theoretical framework and new findings. Psychology of Learning and Motivation, 26, 125–173. Source
Types and levels of automation
What it established. Automation is not on or off. It applies in degrees, and independently across information acquisition, analysis, decision selection, and action, with consequences for the operator that differ by stage.
What it licenses here. Calibrated trust as a graded choice rather than a binary one, and the claim that the same tool used at a different level is a different engagement. It is also why the modes are functions rather than a ladder: a high level of automation on a settled factual question is not a worse choice than a low one.
Parasuraman, R., Sheridan, T. B., & Wickens, C. D. (2000). A model for types and levels of human interaction with automation. IEEE Transactions on Systems, Man, and Cybernetics — Part A, 30(3), 286–297. Source

Supports Behaviors

The organization of turn-taking
What it established. Conversation is locally managed through a describable turn-taking system, which makes what participants are doing analyzable from the record of the talk itself rather than from what they say about it afterwards.
What it licenses here. Coding conversations into modes at all. The turn is the unit and a mode is a characterization of a run of turns. What it does not license is treating a mode label as a fact about the person; it is a description of an episode.
Sacks, H., Schegloff, E. A., & Jefferson, G. (1974). A simplest systematics for the organization of turn-taking for conversation. Language, 50(4), 696–735. Source
Language use as joint action
What it established. Language use is joint action built on common ground: what participants take to be mutually established shapes what any given utterance accomplishes.
What it licenses here. Why the same words are different behaviors depending on what has already been established, and why classification reads runs of turns rather than isolated prompts. A single prompt read out of context is not enough to name a mode.
Clark, H. H. (1996). Using Language. Cambridge University Press. Source

The claim that still has to be earned

The theory predicts that preserving meaningful evaluation should protect later judgment and learning better than delegating evaluation along with production. Existing studies constrain that prediction, but they do not yet establish the complete causal path. Delayed, unaided, and behavioral tests are still required.

Quick check

What distinguishes evaluative agency from simple delegation?

Evaluative agency says what must remain with the person. The next step shows why the right allocation also depends on the person, tool, task, and context.