Pith. sign in

REVIEW 4 major objections 5 minor 6 references

Governing Well in the Algorithmic Age: The Foundations of Digital Statecraft

T0 review · 4 major / 5 minor · reviewed 2026-08-01 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Legitimate government in the algorithmic age has a structural limit: a class of decisions that cannot be delegated to AI no matter how capable it becomes.

desk verdict A useful organizing framework for AI governance, with a central hard limit that is asserted rather than proven. read the letter →

arxiv 2607.18483 v1 pith:LFB77Y6R submitted 2026-07-20 cs.CY cs.AIcs.ETcs.SIcs.SYeess.SY

Zeynep Engin , Tim Gordon , Viviana Bastidas , Tom Crick , Jon Crowcroft , Jean-Martin Denis , David J. Hand , Lauren Maffeo
show 8 more authors
Jakob Mökander Irene Ng Anastasija Nikiforova Giulio Quaggiotto David Uriel Socol de la Osa Rhonda Syler Philip Treleaven Stefaan Verhulst
This is my paper
classification cs.CYcs.AIcs.ETcs.SIcs.SYeess.SY
keywords digitalstatecraftlegitimacyaccountableauthoritynon-delegablecorealgorithmicgovernancehuman-machinecomplementaritysystemiccoherencepublicinterest
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper introduces digital statecraft as the missing conceptual foundation for governing the digital substrate of states—data, algorithms, infrastructure, and platforms—and for governing with those systems. Its central claim is that legitimate governing authority requires two conditions at once: the technical ability to act, and a legitimacy that can justify and account for that action. Those requirements, the paper argues, are in genuine tension, and that tension has a structural limit: the Non-Delegable Core, a class of decisions involving fundamental rights, coercive power, irreversible consequences, or contextual moral judgment that cannot be delegated to algorithmic systems regardless of their capability. From this foundation the paper derives ten principles of digital statecraft, each naming a condition whose absence produces an identifiable governance failure.

What carries the argument

The central object is the Non-Delegable Core (NDC), a class of governing decisions—those involving fundamental rights, coercive power, irreversible consequences, or contextual moral judgment—whose legitimacy depends on human authority being constitutively present in the act of deciding. The NDC is paired with the accountability-capability paradox: the more capable an algorithmic system becomes, the harder it is for humans to second-guess it, which makes the structural requirement of human presence more, not less, necessary. The machinery also includes the over/with distinction (governing over digital systems versus governing with them) and the two-level structure of the NDC (decision-level a

What would settle it

A functioning jurisdiction in which a fully automated decision affecting fundamental rights (e.g., welfare eligibility or parole) has operated for years with unbroken, demonstrable accountability—traceable to a named human authority who accepts responsibility, is subject to sanction, and maintains public confidence—would refute the Non-Delegable Core. Under the paper's account, such a case should not exist.

Watch

Extended reading notes

Core claim

The paper claims that 'digital' does not merely extend the domain of statecraft but reconstitutes it: the machinery of government is being rebuilt around algorithmic systems, non-state actors, and distributed authority. On its account, the state must be able to act on the digital substrate (statecraft over systems) and through it (statecraft with systems), and both dimensions require technical coherence and legitimate authority. The most distinctive claim is that the tension between these requirements has a limit—the Non-Delegable Core—where accountability, traceable authorisation, and judgment across time converge on a requirement that technical capability cannot satisfy: a human agent must

Load-bearing premise

The entire Non-Delegable Core rests on the philosophical premise that authorisation and accountability are structurally human relationships—that a chain of authority ending in a machine cannot be traced to anyone—and on the empirical claim that only the state currently holds the full set of legitimacy conditions.

Editorial extensions

If this is right

  • If the Non-Delegable Core holds, jurisdictions that automate welfare, bail, immigration, or child protection decisions without a human agent constitutively present in the act of deciding are structurally illegitimate, regardless of how accurate the underlying systems become.
  • The over/with framework supplies a diagnostic: governance failures can be located as failures of ability or failures of legitimacy, and the traceability condition demands both horizontal (cross-border) and vertical (devolved-tier) accountability.
  • Institutions must become hybrid—operating at algorithmic tempo with machine monitoring and human judgment—rather than periodic oversight bodies, and adaptive to shifting human-machine configurations.
  • Digital sovereignty is reframed as a legitimacy condition: enforceable accountability over infrastructure, not ownership or territorial control, since traceable authority breaks when states depend on infrastructure they cannot hold to account.
  • The ten principles together name two concrete failure modes: technocratic leviathan (ability without legitimacy) and platform sovereignty (governing authority migrating to companies without democratic mandate).

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The NDC argument, if sound, generalizes beyond states: any entity exercising quasi-public governing authority—including platform companies that shape public discourse—would be bound by the aggregate NDC's requirement of an accountable monitor for systemic effects.
  • A testable implication of the accountability-capability paradox is that as AI models improve, formal human-in-the-loop arrangements will increasingly become nominal rather than substantive; this can be measured by tracking whether human reviewers override system recommendations less over time and whether their reviews are ever acted upon.
  • The framework implies that 'governability by design' should be treated as a procurement requirement, with contractual penalties for ungovernable systems—an operational extension the paper gestures at but does not develop.
  • A pilot comparison could test the theory directly: jurisdictions that adopt constitutionally-present human authority for NDC decisions versus those relying on post-hoc review, measuring accountability outcomes such as successful appeals, sanctions of responsible officials, and public trust.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. This conceptual paper introduces 'digital statecraft' as an organising concept for the governance of and through digital systems. It proposes two foundational requirements — technical coherence (ability) and legitimate authority — and argues that these are in structural tension. From these foundations it derives ten principles, culminating in the Non-Delegable Core (NDC): a class of governing decisions that cannot be legitimately delegated to algorithmic systems regardless of technical capability. The paper also introduces an over/with distinction (statecraft over the digital substrate versus statecraft with algorithmic instruments), develops systems-level and institutional-level ability conditions, and defends a state-centric starting point while holding open the question of whether statehood itself is being reconstituted. The article is explicitly positioned as foundational and agenda-setting, with six research lines left for the field.

Significance. If the NDC claim is correct, the paper supplies a principled boundary for which government decisions cannot legitimately be automated, and the over/with framework plus the vertical/horizontal traceability distinction provide useful diagnostic tools for accountability failures. The manuscript is unusually careful in its hedging: footnotes 4 and 6 concede real limits, and Section 4 candidly states that the framework does not answer whether states can rebuild the required conditions. Its systems-level ability table (Table 2) and the distinction between decision-level and aggregate NDC are genuine contributions. The central limitation is that the paper's most distinctive claim — the NDC as a structural, not contingent, limit — rests on an extra philosophical premise that is asserted rather than defended. The 'derivation' of ten principles is also better described as analytic articulation from chosen definitions than as logical derivation. These are fixable within the manuscript's scope, but they are load-bearing.

major comments (4)
  1. [Section 3.3 and Principle 10] The NDC's hard limit rests on an unargued premise. The Section 3.1 conditions — accountability, traceable authority, consent, revisability — do not by themselves entail that a human agent must be 'constitutively present in the act of deciding.' A chain that runs legislature → minister → algorithm satisfies traceable authority if the minister is genuinely answerable, and revisability can be met by effective ex post review before a human tribunal. The rebuttal to the 'framework authorship' objection shows only that ex ante rule-making is insufficient for underspecified decisions; it does not show why ex post human review cannot bear the required accountability. Robodebt demonstrates the catastrophic consequences of removing human review, but it does not demonstrate a structural impossibility of post-hoc accountability. Consequently, Principle 10 reduces to a demanding but contingent design
  2. [Sections 1 and 4, Table 4] The paper repeatedly says the ten principles are 'derived' from the two foundational requirements, but the derivation step is never shown. Most principles (e.g., Public Interest First, Human Centricity and Civic Agency) are normative assertions introduced with arguments from public-value and democratic theory; they do not follow from technical coherence and legitimate authority alone. The stated test — 'each naming a condition whose absence produces an identifiable and structural governance failure' — is not operationalised, so it cannot distinguish derivation from taxonomy. The authors should either make the derivation explicit or reframe the contribution as an analytic framework with normative commitments, rather than a derivation.
  3. [Section 4] The state-centric premise is load-bearing but asserted: 'The state is the only candidate for which the full set of legitimacy conditions is institutionally available.' No argument is given for why democratic mandate, constitutional accountability, and legal continuity are jointly available only to states, nor for why the state's deficits are all 'contingent — buildable, not inherent.' The paper's own concession that Global South states face 'structural' constraints on infrastructure dependency and regulatory capacity complicates this distinction. If the claim is empirical, evidence is needed; if it is normative, the paper should say so and justify the priority. As written, the conclusion that digital statecraft must begin from the state is stronger than the supporting text warrants.
  4. [Section 3.3 vs. Section 3.1] The NDC is introduced as a claim about 'a class of decisions' but is then split into decision-level and aggregate levels. The aggregate NDC is a different kind of claim: it concerns patterns across decisions, not the act of deciding. This is a useful extension, but the relationship between the two levels is not fully specified. In particular, if aggregate NDC is the same 'underlying test' applied at population level, the paper should explain why an absent monitor violates the Section 3.1 accountability condition even when no individual decision does. As written, the aggregate NDC appears to be an additional normative requirement rather than a consequence of the same derivation.
minor comments (5)
  1. [Note 6 / Section 3.3] The phrase 'authorisation is structurally a human relationship (Hart, 1961)' is a strong interpretive claim. Hart's concept of law concerns rule-governed legal systems, not the claim that authorisation must terminate in a human actor. Either provide a direct quotation or qualify the attribution.
  2. [Section 2.3] The example of a fraud-detection system migrating from the over-axis (procured) to the with-axis (state-operated) is initially clear, but the with-axis example conflates internal accountability structures with the state's relationship to its own systems. The over/with boundary deserves a crisper definitional statement to avoid the appearance of post-hoc classification.
  3. [Section 3.3 and Principle 10] The NDC is cited to Engin & Hand (2025b, Zenodo) and the ten principles to the DSA Founding Manifesto (Engin et al., 2025a). Since the NDC is the article's most distinctive claim, the argument should be self-contained; relying on a non-peer-reviewed preprint and a manifesto is risky for a journal submission.
  4. [Section 4 / Table 4] The text says the principles are 'derived from this tension,' while Section 4 says 'the derivation offered here is foundational rather than exhaustive.' These two framings should be harmonised, especially if the authors decide to move away from derivation language.
  5. [General] Minor copy issues: 'fro m' appears in the running header text; 'for the derivation' in Section 4 should be 'the derivation'; the phrase 'diagnostic and severally correctable' in Section 4 is unclear. These are presentation issues only.

Circularity Check

0 steps flagged · score 2.0 of 10

No formal circularity: the paper's self-citations are attribution and provenance, not load-bearing proof; the NDC's 'constitutively present' requirement is an extra premise, which is a justification gap rather than a circular reduction.

full rationale

This is a conceptual/policy paper with no fitted parameters, no equations, and no empirical predictions, so the fitted-input and definitional-equivalence failure modes do not arise in their usual form. The ten principles are presented as derived from the ability/legitimacy distinction, but the derivation is a normative framework-building exercise rather than a formal derivation; some principles are close to the framework's own stipulations, but that is a property of a conceptual framework, not circularity. The most distinctive claim, the Non-Delegable Core, is attributed to Engin & Hand 2025b, and the ten principles are said to develop the authors' 2025a DSA Manifesto. However, these citations are not the paper's argument: Section 3.3 defends the NDC using the Section 3.1 conditions and external authorities (Hart, Bovens, Pettit), and it explicitly addresses three objections, including the strongest consistency/inspectability objection. The genuinely contested move is from Section 3.1's 'Accountability requires a human agent or institution that bears responsibility... and can be sanctioned, removed, or overruled' to Section 3.3's 'Accountability requires a human agent constitutively present in the act of deciding, not supervising from a distance or reviewing outputs after the fact.' That is not a logical consequence of the earlier definition; it is a stronger normative premise. If unsupported, that is an argumentative gap or correctness risk, not a circularity by construction. Similarly, the assertion that the state is the only actor with the full set of legitimacy conditions is a stipulated starting point, acknowledged in the abstract and conclusion, not a derived empirical result. No load-bearing step reduces to its own input by definition, and the self-citations are not used as the sole justification for the central claims. The paper is therefore not significantly circular, though the NDC's hard limit rests on a contestable premise that the paper does not fully establish.

Assumptions & free parameters 3 free parameters · 5 assumptions · 4 invented entities

For a conceptual framework, the 'free parameters' are hand-chosen conceptual foundations rather than numbers. The framework's output (ten principles, NDC boundary) is determined by these choices plus the axioms listed; readers should treat the framework's scope as conditional on them. The absence of any empirical data means the urgency claims rest on cited literature and asserted trends rather than measurements.

free parameters (3)
  • Foundational pair 'ability' and 'legitimacy' = n/a - qualitative choice
    Hand-chosen conceptual foundation for the entire framework; Footnote 2 acknowledges competing candidates (resilience, sovereignty, representation, trust) are excluded as derivative or constituent. If a different pair were foundational, the ten principles would change.
  • NDC boundary criteria (fundamental rights, coercive power, irreversible consequences, contextual moral judgment) = undefined - deferred
    Determines which decisions fall inside the Non-Delegable Core; the paper explicitly leaves contextual specification to the field (Section 3.3, Footnote 6), so the boundary is a free choice, not a derived result.
  • State-centric starting point = n/a - qualitative choice
    The article 'takes the state as the starting point' (Sections 1 and 4) and asserts it is the only actor with the full legitimacy set; this modeling choice determines the scope of every principle.
assumptions (5)
  • domain assumption Authorisation is structurally a human relationship; only humans can bear responsibility
    Invoked in Sections 3.1 and 3.3 (citing Hart 1961) to ground traceable authority and the NDC. If responsibility can be borne by non-human institutional structures, the NDC collapses. Unproved in the paper and contested in the literature.
  • domain assumption Legitimacy requires answerability to rights-bearing human subjects
    Section 3 opening: 'That party cannot be an aggregate metric... it can only be a rights-bearing subject.' Underpins Principle 7 and all legitimacy conditions.
  • domain assumption The state is the only current candidate institutionally able to satisfy the full legitimacy conditions (democratic mandate, constitutional accountability, legal continuity)
    Section 4; asserted, not derived. Contested by other governance theories; the paper cites Held and Zurn but does not engage them on this point.
  • ad hoc to paper 'Digital' reconstitutes the statecraft question rather than merely extending its domain
    The paper's founding claim (Section 1); defines the stakes of the whole article but is asserted rather than demonstrated.
  • domain assumption Empirical claim that private actors currently set the direction of digital governance
    Section 1: 'The direction of travel is currently being set largely by private actors optimising for commercial metrics...' Asserted with citations but no systematic data; Data Statement confirms no empirical data.
invented entities (4)
  • digital statecraft
    purpose: Organizing concept for the field: governing over and with digital systems under combined ability and legitimacy constraints
    Normative conceptual framing with no falsifiable handle. The term itself appears in earlier cited work (Fourcade & Gordon 2020; Skorpen & Reinertsen 2025); the novelty is in the systematic formulation, not the term.
  • Non-Delegable Core (NDC)
    purpose: Boundary marking decisions that cannot be legitimately delegated to algorithmic systems; operates at decision level and aggregate level
    Normative limit claim with no independent empirical handle; Robodebt and content moderation are retrospective illustrations, not falsifiable predictions. Sourced from the authors' own Engin & Hand 2025b.
  • technical coherence
    purpose: Structural ability requirement uniting systems-level governability, institutional hybridity, and dynamic adaptation
    Re-bundles existing concepts (governability by design, hybrid oversight, adaptive governance) under one heading; a conceptual entity not empirically testable as a unit.
  • aggregate NDC
    purpose: Extends non-delegability to population-level patterns (content moderation, media provenance) where no single decision meets the decision-level threshold
    Conceptual extension of the NDC; a normative claim about accountability of aggregate effects, with cited cases but no predictive test.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Governing Well in the Algorithmic Age: The Foundations of Digital Statecraft." pith.science (2026). https://pith.science/paper/LFB77Y6R

@misc{pith2026260718483,
  author       = {Pith},
  title        = {Pith review of: Governing Well in the Algorithmic Age: The Foundations of Digital Statecraft},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LFB77Y6R}},
  note         = {Machine review of arXiv:2607.18483}
}
read the original abstract

The digital substrate of states -- data, algorithms, infrastructure, platforms, applications -- is being governed without adequate conceptual foundations. The ability and legitimacy required to govern this substrate, and to govern with it, are simultaneously misaligned, contested, and structurally absent. We introduce digital statecraft as the organising concept for this emerging field, arguing that 'digital' reconstitutes the statecraft question rather than merely extending its domain. The concept operates on two dimensions - statecraft over digital systems, concerning the authority and capacity of the state in relation to the digital substrate itself, and statecraft with digital systems, concerning the deployment of algorithmic tools as instruments of governing authority. And it rests on two foundational requirements, technical coherence and legitimate authority, that are genuinely in tension. We derive ten principles of digital statecraft from these foundations, each naming a condition whose absence produces an identifiable and structural governance failure: public interest first, human-machine complementarity, governability by design, systemic coherence, hybrid institutions, adaptive governance, human centricity and civic agency, accountable and traceable authority, judgment across time, and the non-delegable core. This article takes the state as the starting point, the institutional form that developed historically in response to the problem of effective and legitimate public governance, and the only current candidate for which the full set of legitimacy conditions is institutionally available. But the digital statecraft programme holds open a deeper question than just whether states can reform themselves: governing well in the algorithmic age may require rethinking the boundaries, scale, and affiliative basis of statehood itself.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

6 extracted references · 1 canonical work pages

  1. [1]

    & Johnson, S

    Acemoglu, D. & Johnson, S. (2023). Power and progress: Our thousand-year struggle over technology and prosperity. Hachette UK Acharya, D. B., Kuppan, K., & Divya, B. (2025). Agentic AI: Autonomous intelligence for complex goals—A comprehensive survey. IEEE Access, 13 , 18912 -18936. https://doi.org/10.1109/ACCESS.2025.3532853 Anderson, B. (1983). Imagined...

    arXiv 2023

  2. [34]

    P., & Tangi, L

    https://doi.org/10.2139/ssrn.3464203 Nikiforova, A., Rodrigues, M. P., & Tangi, L. (2026). Proactive Public Services in the EU: Readiness, maturity, and pathways to implementation. Publications Office of the European Union, Luxembourg. https://data.europa.eu/doi/10.2760/4554259, JRC146641. NIST AI RMF – National Institute of Standards and Technology. (202...

    arXiv 2026

  3. [78]

    https://doi.org/10.5070/LP61150258 Gillespie, T. (2018). Custodians of the Internet: Platforms, Content Moderation, and the Hidden Decisions That Shape Social Media. Yale University Press. Grimmelikhuijsen, S., & Meijer, A. (2022). Legitimacy of algorithmic decision -making: Six threats and the need for a calibrated institutional response. Perspectives on...

    arXiv 2018

  4. [399]

    https://digitalcommons.law.uw.edu/faculty-articles/640 Cavoukian, A. (2009). Privacy by design: The 7 foundational principles. Information & Privacy Commissioner of Ontario, Canada. Charalabidis, Y., Flak, L. S., & Pereira, G. V. (2022). Scientific foundations of digital governance and transformation. In: Public Administration and Information Technology. ...

    2009

  5. [1147]

    So what if ChatGPT wrote it?

    Cohen, J. E. (2019). Between truth and power. Oxford University Press. Collingridge, D. (1980). The social control of technology. Frances Pinter. Crick, T., Chadwick, A., Strong, C., Jiga-Boy, G., Madsen, J., Burkhardt, J., Roozenbeek, J., Harris, L., Horvath, L., et al. (2025). Deepfakes and media literacy. Government Office for Science, UK Government. h...

    arXiv doi:10.48550/arxiv.1711.01134 2019

  6. [1249]

    https://openscholarship.wustl.edu/law_lawreview/vol85/iss6/2 Coeckelbergh, M. (2023). Democracy, epistemic agency, and AI: political epistemology in times of artificial intelligence. AI and Ethics, 3(4), 1341-1350. https://doi.org/10.1007/s43681-022-00239-4 Coglianese, C., & Lehr, D. (2016). Regulating by robot: Administrative decision making in the machi...

    work page doi:10.1007/s43681-022-00239-4 2023

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.