Normative Ethics, Artificial Intelligence, and Value Alignment (Dynamic Normativity)

Nicholas Kluge Corrêa · 2026 · Springer, International Library of Ethics, Law and Technology vol. 26 (based on Bonn/PUCRS cotutelle PhD dissertation) interlocutor high priority coded

Main argument

Thesis (from front matter + structure): 'Dynamic Normativity' - a framework holding that if a particular set of requirements is met, 'a minimal level of alignment has been achieved'; Part I foregrounds the field's history (2015-2020 blog-post era to market-driving mainstream), the author's assumptions and biases; Part II defends the necessity and coherence of the conditions underlying Dynamic Normativity (Ch 4) and offers 'a minimal set of strategies and methodologies for tackling the value alignment problem' (Chs 5-7) with practical artifacts (dashboards, code repositories, demos). Positioning per Markus Gabriel's foreword: articulating 'a shared normative grammar in a technologically mediated world' with 'empirical clarity about the global moral landscape of AI, showing both where consensus exists and where critical reflection must continue.'

Why it matters here

A PUBLISHED PhD DISSERTATION on exactly the dissertation's topic (normative ethics + value alignment), from a Bonn/Brazil cotutelle - the single most relevant structural model for Augustine's own project: how a philosophy PhD on alignment gets organized, examined, and published. Its 'Dynamic Normativity' framework (conditions for minimal alignment + practical toolkits) is also a rival account to position against. Foreword by Markus Gabriel.

Reading notes

Targeted treatment (217pp monograph): foreword, preface, chapter map read; full close read of Chs 4-7 (the Dynamic Normativity conditions + practical strategies) deferred to a dedicated session - PRIORITY for deep reading before the proposal presentation, as this is the closest existing dissertation-shaped competitor/model. Structure: Part I (foundations, Chs 1-3, incl. assumptions/biases foregrounding), Part II (Ch 4 defends the necessity/coherence of Dynamic Normativity's conditions; Chs 5-7 give minimal alignment strategies with dashboards/repos/demos as companion artifacts). Global-South authorship (PUCRS Brazil + Bonn; CAPES/DAAD funded) - relevant to the non-Western thread. DEEP-READ COMPLETED (Chs 4-7 + closing): Dynamic Normativity = three sufficiency requirements - AGGREGATE human preferences coherently under uncertainty, LEARN from them, MITIGATE unwanted behavior/impact - each chapter operationalized (Ch5 preference learning incl. SFT/RLHF/DPO; Ch7 impact mitigation at training and inference time). Metaethical base: epistemic COHERENTISM (Sayre-McCord; possibly Haack foundherentism), deliberately not a substantive moral theory; aggregation restricted to ORDINAL preference sets precisely to dodge intertheoretic comparison problems (MacAskill machinery discussed and sidestepped); criteria: completeness, transitivity, Kolmogorovian weights, Pareto efficiency etc. Ch1 is a large empirical survey (WAIE - Worldwide AI Ethics, 200 guidelines analyzed) with marginalized-regions critique.

Kluge Corrêa, N. (2026). Normative Ethics, Artificial Intelligence, and Value Alignment. Springer.

Close reading — 5 coded units

#1 · pp. 1 · claim

“[Dynamic Normativity: the final three chapters] offer a minimal set of strategies and methodologies for tackling the value alignment problem, arguing that if a particular set of requirements are met, we can say that a minimal level of alignment has been achieved.”

TU-METAETH

#2 · pp. 1 · claim

“[Foreword, M. Gabriel:] his work is not only a contribution to AI ethics but also to a renewed cosmopolitanism of reason: an attempt to articulate a shared normative grammar in a technologically mediated world. [...] it provides empirical clarity about the global moral landscape of AI, showing both where consensus exists and where critical reflection must continue.”

TU-LITREV

#10 · pp. 85–86 · definition

“In Dynamic Normativity, we must satisfy three distinct requirements for a sufficient alignment condition. [...] (1) coherently aggregate human preferences, (2) learn from them, and (3) mitigate unwanted behavior. [...] Aligned AI systems should coherently aggregate human preferences in a way that resolves cases of uncertainty.”

TU-METAETH

#11 · pp. 86–87 · argument

“it is impossible to compare ordinal preference sets with cardinal utility functions (deontological and consequentialist theories). [...] if we cannot extract a choice-worthiness value from ordinal theories [...] it is unclear how to use them in cases of uncertainty. [...] [and] the problem of intertheoretic comparisons. [...] To address the problems above, we will focus on aggregating human preferences expressed as ordinal sets in this study.”

NF-PLURAL-OTHER

#12 · pp. 87 · argument

“this aggregation requirement is not rooted in any quintessential metaethical blueprint but in a comprehensive and desirable set of criteria for aggregating preferences. [...] these criteria can be inspired by an epistemic Coherentist view (Sayre-McCord 1996), i.e., the idea that epistemically justifiable methods should be part of any process related to dealing with uncertainty (moral or empirical) and forming knowledge.”

TU-MWRE

Synthesis-matrix row

complicates T8-NONWESTERN-CONCEDED

Global-South authorship; cosmopolitan framing

Memos (2)

thesis-link · unit #1

STRUCTURAL MODEL ALERT: this is a philosophy PhD dissertation (Bonn + PUCRS cotutelle, begun 2021, defended, then published in a major Springer ethics series with a Markus Gabriel foreword) on normative ethics + value alignment with EMPIRICAL components and practical artifacts (dashboards, repos - exactly Augustine's dashboards). It is simultaneously: (a) proof to Howard/the department that this dissertation-shape is examinable and publishable; (b) the closest competitor to differentiate from - Kluge Corrêa's consensus-mapping ('where consensus exists') is adjacent to convergentism but appears not to have the Rossian object-level theory, the responsibility-attribution focus, or the stakeholder-stratified corpus; (c) a checklist source - before the proposal presentation, deep-read Chs 4-7 to confirm the differentiation claims. HIGH PRIORITY follow-up read.

theoretical · unit #11

DIFFERENTIATION CONFIRMED (deep read): Kluge Corrêa and the dissertation share two formal moves - ordinal-only treatment (his unit 11 = the dissertation's ordinal-convergence reply to Lloyd) and a coherentist epistemology (his Sayre-McCord base = the dissertation's MWRE bridge). The differences are now precise: (1) OBJECT-LEVEL THEORY - Dynamic Normativity deliberately avoids any substantive moral theory ('not rooted in any quintessential metaethical blueprint'), remaining a procedural-epistemic account of PREFERENCE aggregation; the dissertation runs substantive frameworks (Ross/consequentialism/contractualism) and uses their CONVERGENCE - so Kluge Corrêa remains within what Zhi-Xuan would call the preferentist data model, while the dissertation works with reasons/values; (2) RESPONSIBILITY - no allocation theory anywhere in Dynamic Normativity (mitigation is about system impact, not human accountability); (3) EMPIRICS - his empirical layer is the WAIE guidelines survey (documents), the dissertation's is naturalistic folk discourse at scale (stakeholders); (4) his 'sufficient minimal alignment' is a design standard, the dissertation's convergentism is a justification standard. Complementary more than rival; cite as the nearest neighbor and state these four contrasts explicitly in the intro. Also mine Ch1's WAIE dataset (200 guidelines, marginalized-regions analysis) as a companion empirical source for the governance chapter.