Ethics of Artificial Intelligence

Vincent C. M\"uller

REVIEW 3 major objections 3 minor 1 references

Ethics of Artificial Intelligence

T0 review · 3 major / 3 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read This survey organizes AI ethics around one divide: AI as tool versus AI as moral agent.

desk verdict A standard AI ethics survey, unverifiable from the corrupted text; the abstract is coherent but has a missing section 2.5. read the letter →

arxiv 2508.16658 v1 pith:I3UYH57Q submitted 2025-08-20 cs.CY

Vincent C. Müller This is my paper · ORCID

classification cs.CY

keywords AIethicsmachineartificialmoralagencyalgorithmicbiasopacityprivacyautonomyandresponsibilitysingularity

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

The paper sets out to introduce the main debates in AI ethics for a broad audience and claims that the field falls into two large groups: ethical questions about AI systems as objects, meaning tools designed and used by humans, and ethical questions about AI systems as subjects, meaning systems to which ethics is applied directly. Across the object side it covers privacy, manipulation, opacity, bias, autonomy and responsibility, and the singularity; across the subject side it covers machine ethics and artificial moral agency. For every theme it explains the general problem, surveys existing positions, connects them to current technology, and draws policy consequences. A reader comes away with a structured map that shows where each familiar AI controversy sits and what kind of response each one tends to require.

What carries the argument

The organizing device is the object/subject distinction. AI as object means the system is a tool, so ethics examines the human choices embedded in its design and use; AI as subject means the system is itself a bearer of ethical evaluation, so ethics asks what the machine should do or whether it can be a moral agent. This distinction carries the argument by determining which themes belong together, which arguments apply, and which policy levers follow.

What would settle it

A single well-established AI ethics debate that cannot be assigned to either side without stretching the category—for example, systemic labor displacement or AI's environmental footprint—would show the organizing distinction is incomplete. More directly, a survey of current AI ethics syllabi could check whether the field's dominant topics actually match the list in this paper.

Watch

Extended reading notes

Core claim

In its own terms, the paper's central claim is that the sprawling field of AI ethics can be organized by a single role distinction. When AI systems are treated as objects—instruments made and used by humans—the ethical questions are about how such tools affect people: privacy, manipulation, opacity, bias, responsibility for autonomous behavior, and the long-term risk of a singularity. When AI systems are treated as subjects, the ethical questions concern what the systems themselves should do, giving rise to machine ethics and debates over artificial moral agency. The paper argues these two postures generate different problems and different policy consequences, and it uses the distinction to

Load-bearing premise

Everything rests on the assumption that the topics chosen here are indeed the main debates of AI ethics and that the object/subject distinction is a useful way to sort them; if a major debate falls outside both sides, the overview's map would mislead readers about where the field's center of gravity lies.

Editorial extensions

If this is right

If the map is right, the ethics of AI is not one debate but two families, and conflating them risks category errors: concerns about bias and opacity are tool-use concerns, while machine ethics concerns what the system itself should do.
Policy consequences follow per theme: privacy and manipulation call for regulation of data practices and interface design; opacity and bias call for transparency and fairness requirements; autonomy and responsibility call for accountability rules; the singularity calls for control and safety measures.
Artificial moral agency is treated as a coherent topic only if one grants that systems can be subjects of ethical evaluation, which the paper presents as a live debate rather than a settled fact.
The future shape of the field depends on the concept of AI itself; if that concept shifts, the list of main debates shifts with it.

Reading between the lines

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

The object/subject divide predicts where existing AI ethics frameworks will feel strained: hybrid cases such as human-AI teams or delegated authority sit between the two poles and may call for a third category.
A testable extension would be to classify a corpus of AI ethics cases by this distinction and check whether the categories predict which policy instruments researchers propose.
If accepted, the taxonomy implies that debates about labor displacement and environmental impact are not obviously contained by either side, since they concern systemic consequences rather than tool-use harms or machine agency.

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.

Desk Editor's Note

read the letter

First, the practical issue: the full text we received is encoding-corrupted, so almost nothing in the body is readable. What survives is the abstract, which reads like a textbook table of contents for AI ethics. If the body actually delivers that structure, this is a competent survey in a crowded field—nothing new, but potentially a useful reference for students and people entering the area. Müller is a recognized name, so the citation pattern is probably fine, though I can't confirm from what I have.

The paper's framing is the standard object/subject distinction: AI as tools (privacy, manipulation, opacity, bias, autonomy, singularity) and AI as agents (machine ethics, moral agency). There's no new argument or framework. The novelty is organizational, and that's okay for a survey, but it means the bar for evaluation is accuracy, completeness, and clarity—none of which I can actually check.

The soft spots are proportional to what I can see. First, the abstract itself contains a numbering glitch: it lists sections 2.1 through 2.4, then jumps to 2.6 and 2.7, with no 2.5. That could be a typo in the abstract or a missing section in the body; I can't tell. Second, the abstract's claim to cover the 'main debates' is assertive. It omits, for example, AI and labor or environmental impact, which some would argue are central. That's a potential criticism, but the abstract doesn't promise exhaustiveness, so it's not a fatal flaw. Third, and most important, the unreadable text means the survey could be accurate or misleading—I have no way to judge.

Who is this for? Someone who wants a quick map of the canonical topics in AI ethics. Not for specialists or for someone looking for a novel thesis. If the actual article is intact, it's the kind of thing I'd assign to incoming grad students. It deserves a serious referee only insofar as any encyclopedia entry does—there's no research claim to scrutinize, but there is the question of whether the overview is fair and well-organized.

My recommendation: don't treat this arXiv version as the final word. Ask the author for a clean copy or find the published version. If the body matches the abstract and the section numbering is fixed, it's a perfectly citable survey. As it stands, I can't endorse or reject.

Referee Report

3 major / 3 minor

Summary. This manuscript is a survey article on the ethics of AI. Its abstract presents a structured roadmap: after background (Section 1), it distinguishes AI systems as objects of ethical analysis (privacy 2.1, manipulation 2.2, opacity 2.3, bias 2.4, autonomy & responsibility 2.6, singularity 2.7) from AI systems as subjects (machine ethics 2.8, artificial moral agency 2.9), and closes with future developments and the concept of AI (Section 3). For each theme the abstract promises a general explanation of the ethical issues, existing positions and arguments, analysis in light of current technologies, and policy consequences. The abstract is coherent and standard for an introductory survey. However, the full text supplied for review is heavily corrupted with encoding errors; only the abstract and some disconnected fragments are readable. Consequently, the actual body of the survey, its accuracy, coverage, and citations, could not be verified from the provided artifact.

Significance. If the body fulfills the abstract's promise, this would be a useful structured introductory survey for non-specialists, with a sensible organizing distinction between AI as object and AI as subject, and with the added value of policy implications per theme. The transparency of the proposed structure is a strength, and the paper would be pedagogically serviceable. However, the significance is conditional: because the body text is unreadable in the version under review, I cannot confirm that the named debates are accurately represented, that the promised per-section analysis actually exists, or that the coverage is balanced. The contribution can only be evaluated after a readable version is available.

major comments (3)

[Abstract, Section 2 enumeration] The abstract lists the components of Section 2 as 2.1 privacy, 2.2 manipulation, 2.3 opacity, 2.4 bias, then 2.6 autonomy & responsibility, 2.7 singularity, 2.8 machine ethics, and 2.9 artificial moral agency. There is no Section 2.5. This is a concrete internal inconsistency in the paper's roadmap. It may be a typographical slip or it may indicate a missing section, but either way it must be corrected because readers will rely on the abstract to navigate the survey.
[Full text] In the manuscript version provided for review, the full text is largely unreadable, consisting mostly of replacement characters. Only the abstract and some fragments can be parsed. This is a load-bearing limitation: the paper's central claim is that it 'introduces the main debates' and covers each theme with explanation, positions, technology analysis, and policy consequences. None of this can be checked in the current artifact. Sections 2.1–2.9 and Section 3 are not accessible. The authors should submit a clean, readable version so the body can be assessed; as it stands, the soundness of the survey cannot be verified.
[Abstract, scope claim] The abstract asserts that the listed topics are 'the main debates' in AI ethics without giving selection criteria. The list omits several areas that are commonly treated as central in the field, such as AI and labor, environmental impacts, and existential risk beyond the singularity. If these are also absent from the body, the claim to cover the main debates is too strong. The authors should either add such debates or qualify the claim (e.g., 'some main debates' or 'a selection of central debates').

minor comments (3)

[Abstract, punctuation] The numbering '2.8.' and '2.9.' includes a period after the section number, while other entries (2.1–2.4, 2.6, 2.7) do not. Please regularize the style.
[Abstract, readership] The abstract does not state the intended audience or assumed background. A sentence indicating that the survey is aimed at students or non-specialists would help readers calibrate expectations.
[General] The phrase 'autonomy & responsibility' mixes a symbol and a word; consider writing 'autonomy and responsibility' for consistency.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the paper is a survey article with no derivation or fitted prediction chain to reduce.

full rationale

This manuscript is a survey of AI ethics debates, not a derivation or prediction paper. It makes no formal claims, fits no parameters, and presents no equations whose outputs could equal their inputs by construction. The central claim—that the article "introduces the main debates" under an object/subject distinction—is an organizational assertion about the literature, not a result derived from a self-referential premise. The provided full text is heavily corrupted, making the body unverifiable, and the abstract's section numbering skips from 2.4 to 2.6, but these are readability and structural-accuracy concerns, not circularity. No load-bearing self-citation, uniqueness theorem, ansatz smuggled by citation, or renaming of a known result can be identified from the available artifact. Per the rule that a non-finding is expected when warranted, the circularity score is 0.

Assumptions & free parameters 0 free parameters · 2 assumptions · 0 invented entities

A review paper introduces no free parameters or invented entities. The only structural assumptions are the selection and categorization of topics, which are standard domain assumptions for an overview article.

assumptions (2)

domain assumption The selected topics (privacy, manipulation, opacity, bias, autonomy and responsibility, singularity, machine ethics, artificial moral agency) constitute the main ethical debates in AI.
The abstract presents this list as the 'main debates' without argument; the survey's coverage depends on this choice.
domain assumption AI systems can be meaningfully analyzed as objects (tools) and as subjects (moral agents).
The abstract organizes the entire survey around this dichotomy, which is asserted rather than argued.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Ethics of Artificial Intelligence." pith.science (2026). https://pith.science/paper/I3UYH57Q

@misc{pith2026250816658,
  author       = {Pith},
  title        = {Pith review of: Ethics of Artificial Intelligence},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/I3UYH57Q}},
  note         = {Machine review of arXiv:2508.16658}
}

read the original abstract

Artificial intelligence (AI) is a digital technology that will be of major importance for the development of humanity in the near future. AI has raised fundamental questions about what we should do with such systems, what the systems themselves should do, what risks they involve and how we can control these. - After the background to the field (1), this article introduces the main debates (2), first on ethical issues that arise with AI systems as objects, i.e. tools made and used by humans; here, the main sections are privacy (2.1), manipulation (2.2), opacity (2.3), bias (2.4), autonomy & responsibility (2.6) and the singularity (2.7). Then we look at AI systems as subjects, i.e. when ethics is for the AI systems themselves in machine ethics (2.8.) and artificial moral agency (2.9). Finally we look at future developments and the concept of AI (3). For each section within these themes, we provide a general explanation of the ethical issues, we outline existing positions and arguments, then we analyse how this plays out with current technologies and finally what policy consequences may be drawn.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

1 extracted references · 1 canonical work pages

[1]

�� ...

work page Pith review arXiv 2025

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.

[1] [1]

�� ...

work page Pith review arXiv 2025