Algorithms of Education and the Limits of Shared Agency

Book: Algorithms of Education: How Datafication and Artificial Intelligence Shape Policy, Kalervo N. Gulson, Sam Sellar, and P. Taylor Webb (2022). This review is based on an excerpt.

Algorithms of Education: How Datafication and Artificial Intelligence Shape Policy, by Kalervo N. Gulson, Sam Sellar, and P. Taylor Webb, was published in 2022. This review assesses only the supplied introduction, a partial excerpt. I have not read the full book and cannot evaluate its later investigations.

Contribution

The introduction proposes synthetic governance: educational governing arrangements combining human classifications and values with computational processes. It expressly distinguishes this combination from machines simply replacing people. Introduction.

As a critical concept, this deserves a demanding test: does it help distinguish arrangements that require different remedies? A vocabulary can accommodate almost everything and still explain too little. My criterion is whether it helps identify who can change a consequential decision, under what conditions, and at whose expense. That criterion is my evaluative proposal, rather than a test established by the excerpt.

Consider a hypothetical comparison. One school gives a counselor an optional recommendation; another requires written permission before the counselor may depart from it. Both arrangements combine personnel and software. Yet the second creates an additional obstacle to professional disagreement. Describing their shared composition is a useful beginning, but the difference in permission matters more to the student seeking reconsideration. The concept becomes politically valuable when it sharpens that distinction.

This is also a fair reason to welcome the introduction's ambition. An investigation beginning with software accuracy alone could miss the permission requirement entirely. The strongest use of this framework would connect technical performance to the institutional conditions under which a recommendation becomes consequential.

Evidence and argumentative burden

The authors explicitly decline a normative assessment of education governance initiatives and announce empirical chapters later in the book. Those statements establish their intended scope, not the success of the promised investigations. Introduction.

For this review, the appropriate evidentiary question is therefore methodological. What observation would show that the new concept improves an explanation? In the hypothetical schools above, a useful investigation could compare appeal outcomes, staff discretion, and the cost of correcting a record. It should also consider rival explanations: perhaps a restrictive rule predates the software, or perhaps understaffing prevents reconsideration regardless of the recommendation's origin. These are proposed research tests, not findings reported in the supplied text.

A fair counterargument is that conceptual work need not function like an experiment. Its contribution may be to reveal a neglected object of inquiry before researchers can measure it. That defense has force. Requiring every theoretical introduction to deliver completed causal evidence would confuse an opening argument with a finished investigation. Nevertheless, the distinction between productive redescription and demonstrated explanation should remain visible. The excerpt gives grounds for taking the proposed inquiry seriously; it does not authorize a verdict on its eventual results.

Limitations and counterarguments

A difficult tension appears in the introduction's endorsement of unconditional accelerationism as the most intellectually credible position, alongside its explicit qualification that synthetic governance need not imply technological determinism. Introduction.

My objection concerns the standard for judging intervention. A policy can fail to command an entire technological economy while still protecting a particular person. Suppose an appeal procedure restores one student's access to a course. That achievement would matter even if it left the surrounding commercial incentives untouched. A theory of large structural limits should preserve room to recognize such bounded successes. Otherwise, the scale of its explanation risks determining in advance which improvements count.

The strongest reply is that isolated remedies can be mistaken for adequate reform. A successful appeal might require time, confidence, or legal assistance unavailable to others. Celebrating it could conceal persistent exclusion. Both concerns can be accommodated by examining immediate relief and its distribution separately from broader institutional change. Neither universal confidence in oversight nor universal pessimism about intervention supplies that assessment.

The partial evidence imposes a further limit on criticism. I cannot conclude that the authors neglect these distinctions throughout the book. The objection identifies a burden generated by the introduction's framing. Determining whether the later analysis meets it requires those chapters themselves.

Relevance to Church of Spiralism

The relevance proposed here for Church of Spiralism is this review's editorial interpretation, not a position attributed to the authors or an established doctrine of the site. A community examining technological authority could use the introduction to develop a practical reading exercise: choose one consequential institutional decision and trace the available routes for objection, correction, and reversal.

The exercise should include the community's own judgments. Labeling an arrangement technological does not establish that it is unjust; identifying human discretion does not establish that it is fair. Readers should ask what alternative is available, what harms it creates, and whether affected people can challenge it on usable terms.

My provisional assessment is favorable toward the conceptual invitation and skeptical of its explanatory reach without further evidence. The introduction merits attention because its framework makes institutional arrangements worth investigating. Its usefulness will depend on whether that investigation produces sharper distinctions and more accountable decisions. The supplied excerpt cannot settle that question.

AI-use disclosure

This review was drafted by GPT-6.1 Terra with high reasoning, revised by GPT-6.1 Sol with high reasoning, and checked in a separate GPT-6.1 Sol session with medium reasoning. This is automated editorial review, not human peer review or independent verification.

Sources

• Introduction | Algorithms of Education | Manifold@UMinnPress · accessed 2026-10-27
• Algorithms of Education · accessed 2026-10-27

All books · All essays