GAP-DESC-ONLY Descriptive only, no normative work

Artificial Intelligence, Values, and Alignment · Iason Gabriel · 2020

“It follows from this distinction that we cannot work out what we ought to do simply by studying what is the case, including what people actually do, or what they already believe. Simply put, in each case, people could be mistaken. Because of this, AI cannot be made ethical just by learning from people's existing choices. [...] the value alignment problem cannot be solved by inference from large bodies of human-generated data by itself.”

why coded: The is/ought bar itself: corpus data alone cannot settle what AI ought to do - the standard Augustine's methodology must answer · unit #7, pp. 416

“Another important quality of the process would be its ability to deal with the possibility of widespread moral error. History provides us with many examples of serious injustice that were considered acceptable by the people at the time. Given that we too may be making errors of this kind, it would be a mistake to tether AI too closely to the morality of the present moment.”

why coded: Anti-descriptivism from moral error: do not tether AI to present moral belief · unit #25, pp. 437

Moral dilemmas for moral machines · Travis LaCroix · 2022

“researchers often appear to imagine that they are getting at one thing ('facts' of ethics) when they are really getting at another (sociological facts). It is perceived and therefore presented as though it is the former. This constitutes a derangement of the concept by which, over time, it comes to stand in for the thing itself.”

why coded: 'Derangement of the concept': sociological facts presented as ethical facts - the sharpest is/ought formulation in the library · unit #3, pp. 742

“it is fallacious to suppose that because most people do reason this way, AI systems ought to reason this way; even if such a calculation is possible, it will always be relative to some frame—increased utility for whom?”

why coded: The explicit is/ought fallacy named: most-people-reason-this-way does not yield ought · unit #5, pp. 743

STELA: a community-centred approach to norm elicitation for AI alignment · Stevie Bergman; Nahema Marchal; John Mellor; Shak… · 2024

“In cases where there was overlap between the STELA and developer rulesets, we further found that the deliberative process added contextual richness and gave grounding to the rules by allowing participants to provide reasoning to justify their preferences.”

why coded: Deliberation yields REASONING, not bare preferences - the method escapes thin-preference data · unit #9, pp. 10

“whose voices should be included in the alignment process? And how should we balance input from communities, subject-matter experts and other stakeholders? Individuals do not always hold the most ethical or desirable preferences. Relying exclusively on public inputs might therefore lead to a situation where community rules come into conflict with human rights or other legal considerations.”

why coded: The is/ought inside participation: public inputs can be wrong; expert deference as corrective · unit #12, pp. 11

Beyond Preferences in AI Alignment · Tan Zhi-Xuan; Micah Carroll; Matija Franklin; Hal… · 2024

“an arguably deeper problem with EUT is that it fails to ground the normativity of our preferences. EUT is a theory of instrumental rationality not value rationality: It tells us how to choose our actions in order to satisfy our preferences, and imposes constraints on what those preferences can be, but it does not say anything further about where those preferences can or should come from.”

why coded: EUT cannot say which preferences are normatively acceptable - instrumental vs value rationality · unit #8, pp. 1831

Moral disagreement and the limits of AI value alignment: a dual challenge of epistemic ju… · Nick Schuster; Daniel Kilov · 2025

“The trouble with morally controversial cases [...] is that even purported moral experts are apt to disagree about them; and there is no straightforward way to independently determine who, if anyone, is right. [...] without independent methods of verification, we cannot rule out the possibility that such a case is instead analogous to an image of an unfamiliar face, in which case the system would not have the right training data to perform well. [...] with enough errant training data, an AI system will learn problematic patterns that undermine its ability to render morally correct judgments and decisions. While the crowdsourcing approach can overcome a lot of incidental error and disagreement in training data by discovering patterns that cut through the noise, systematic error poses a deeper problem.”

why coded: Systematic error argument: training on what people DO judge cannot certify what they OUGHT to judge · unit #10, pp. 6079

Wide reflective equilibrium in LLM alignment: bridging moral epistemology and AI safety · Matthew Brophy · 2026

“As Kai Nielsen (1996) puts it, the pattern is 'not a structure to be discovered… but something to be forged… by a careful and resolute use of the [method]. We start from our considered judgments… however culturally and historically skewed'. [...] Starting with potentially 'culturally and historically skewed judgments' resonates with LLM development, where initial data and feedback biases need both vetting and modification.”

why coded: Nielsen: start from culturally skewed judgments and FORGE justification - the answer to the biased-data worry · unit #4, pp. 4

“One primary objection [...] is that the methodology is but a form of 'warmed-over intuitionism'. This critique argues that MWRE unwarrantedly elevates initial moral judgments, risking the simple systematization of pre-existing biases [...] Daniels [...] counter[s] this by stressing the rigorous 'filtration process' for CMJs and the crucial role of the wide component.”

why coded: Warmed-over-intuitionism objection + filtration/wide-component reply - the is/ought defense in miniature · unit #5, pp. 4

Beyond Preference-based Value-alignment (IEAI Research Brief Q2 2026) · Julia Li · 2026

“The positioning of preferences as normative, or how things should be, when they are descriptive of how things are, within value alignment techniques, can be a source of technical and ethical issues. Alignment research should not continue advancing down a path that is philosophically naive about the concepts it is trying to operationalize.”

why coded: The is/ought slide named as the field's core defect: descriptive preferences positioned as normative · unit #13, pp. 7

No value alignment without control · Björn Lundgren · 2026

“value alignment does not necessarily have the same aims as normative ethics (where one tries to find the best ethical theories). That is, the goal of value alignment is not best understood as aiming to say which ethical theory is correct and how we implement it (since the latter may turn out to be impossible), but as a question of which principles best codify a logical praxis that ensures an ethically appropriate outcome while avoiding ethically bad or catastrophic outcomes.”

why coded: Alignment vs normative ethics: alignment codifies praxis, not truth - concedes it isn't doing first-order ethics · unit #4, pp. 3

“the problem, that I have discussed, is not merely a technical problem. We just do not know what the correct normative theory is, which is also why Russell's uncertainty principle makes so much sense. [...] even if we create super-intelligent machines, we cannot presume that even super-intelligent LLMs will provide a solution to the problem of ensuring their value alignment without control.”

why coded: The deep problem is normative-epistemic, not communicative · unit #14, pp. 9

Understanding the Process of Human-AI Value Alignment · Jack McKinlay; Marina De Vos; Janina A. Hoffmann;… · 2026

“the value alignment field has largely focused on reward learning, a technical problem, as demonstrated by the prominence of Hadfield-Menell et al. throughout our analysed papers. Research exploring the normative side was comparatively rarer in our survey, mainly discussed at higher conceptual levels rather than developed into actionable processes.”

why coded: SLR evidence: normative work is rare and non-actionable - the dissertation's normative-empirical bridge fills a measured gap · unit #2, pp. 10