The DOI matches its Crossref record; 87 of the 89 references it cites were found in a registry or an index; 7 of the 53 sources its claims cite were checked for support.
Four Responsibility Gaps with Artificial Intelligence: Why they Matter and How to Address them
Open accessAbstract
The notion of “responsibility gap” with artificial intelligence (AI) was originally introduced in the philosophical debate to indicate the concern that “learning automata” may make more difficult or impossible to attribute moral culpability to persons for untoward events. Building on literature in moral and legal philosophy, and ethics of technology, the paper proposes a broader and more comprehensive analysis of the responsibility gap. The responsibility gap, it is argued, is not one problem but a set of at least four interconnected problems – gaps in culpability, moral and public accountability, active responsibility—caused by different sources, some technical, other organisational, legal, ethical, and societal. Responsibility gaps may also happen with non-learning systems. The paper clarifies which aspect of AI may cause which gap in which form of responsibility, and why each of these gaps matter. It proposes a critical review of partial and non-satisfactory attempts to address the responsibility gap: those which present it as a new and intractable problem (“fatalism”), those which dismiss it as a false problem (“deflationism”), and those which reduce it to only one of its dimensions or sources and/or present it as a problem that can be solved by simply introducing new technical and/or legal tools (“solutionism”). The paper also outlines a more comprehensive approach to address the responsibility gaps with AI in their entirety, based on the idea of designing socio-technical systems for “meaningful human control", that is systems aligned with the relevant human reasons and capacities.
Overview
CoChat Summary
TL;DR
The paper argues that the AI responsibility gap is not one problem but at least four interconnected gaps—culpability, moral accountability, public accountability, and active responsibility—arising from technical, organisational, legal, ethical, and societal sources. It critiques fatalist, deflationist, and solutionist responses and sketches a comprehensive approach based on designing sociotechnical systems for meaningful human control.
Key findings
- The responsibility gap should be reframed as at least four interconnected problems: gaps in culpability, moral accountability, public accountability, and active responsibility; these can occur with non-learning systems too and have multiple sources, some technical, organisational, legal, ethical, and societal.
- Current partial approaches distort the problem: “fatalism” treats it as new and intractable; “deflationism” dismisses it as not new and not a problem; “solutionism” treats it as solvable by simply introducing new technical and/or legal tools, with risks of “technical solutionism” and “legal solutionism.”
- The paper proposes meaningful human control (MHC) as a more comprehensive design-oriented approach, with two major conditions: “tracking” (the sociotechnical system responds to relevant human reasons and facts) and “tracing” (for each relevant action, at least one human agent along the design, development, and use chain has sufficient knowledge and moral awareness/capacity).
- It illustrates the gaps and proposed approach with examples including automated driving systems, AI-driven medical diagnosis, algorithmic public administration, and the Uber automated vehicle accident.
Methods
- Conceptual/normative analysis and critical literature review in moral and legal philosophy and ethics of technology; no empirical data collection or sample is reported.
- The analysis draws on philosophical, legal, and sociological theories of responsibility (e.g., Hart, Bovens, van de Poel, Gardner) to distinguish four forms of responsibility and map corresponding gaps and causes.
- It uses illustrative cases—automated driving, medical diagnosis, public administration, and the Uber accident—to show how gaps arise and how MHC tracking/tracing might address them.
Limitations
- Stated: the MHC proposal is a sketch; future research will develop it in more detail, and future philosophical, empirical, and technical work must clarify which responsibility gaps emerge in different systems and contexts and the extent to which MHC may effectively address them.
- Evident: the paper is conceptual and does not empirically test or operationalize MHC; its examples are illustrative rather than a systematic empirical sample.
- Evident: the critical review is described as a “representative sample” of the literature, not a systematic review.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
6 Valid · 3 Not valid · 27 Not checked · 62 Own results and interpretations
The cited sources that could be checked couldn’t confirm or dispute this claim.
In the paper
“Moral accountability also has an instrumental value.”
What the sources say
- Not checked — no full text of the sourcecites Collingridge, 1980 source
- Could not verifycites Genus and Stirling (2018) sourceExperimental — accuracy not yet measured
The Introduction discusses ethical responsibility and accountability, but does not say that moral accountability has instrumental value. The Conclusion contrasts responsibility as a struggle against incumbent power with instrumental facilitation; it does not establish or explicitly deny the claimed value. Searches for “moral accountability” and “instrumental value accountability” found no passage addressing the claim. (LLM, conf=0.72)
- Not checked — no full text of the sourcecites Stilgoe et al., 2013 source
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Every source the paper cites for this claim was checked and supports it.
In the paper
“Most of these discussions point to the fact that algorithmic decision-making is often difficult to understand for - and explain to - human agents, due to the different, and sometimes inscrutable for persons, ways of AI operations: the so-called 'black box' problem (Castelvecchi, 2016).”
What the sources say
- Supportedcites Castelvecchi, 2016 source
The unlabelled opening passage describes neural networks as opaque and says their stored information is exceedingly difficult to decipher. The “ILLUSTRATION BY SIMON PRADES” passage explicitly calls this the black-box problem and says it can be difficult for humans to understand how such systems work or for the systems to explain themselves. The source discusses AI, particularly deep-learning neural networks, rather than every algorithmic decision system. (LLM, conf=0.93)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Nor does this approach incentivise engineers to go beyond the current state of art in technology and look for innovative solutions which may improve the capabilities of current technology to better satisfy complex and potentially conflicting societal demands (Van den Hoven et al., 2012) such as, for instance, higher levels of safety combined with better predictability etc.”
What the sources say
- Partly supportedcites Van den Hoven et al., 2012 sourceExperimental — accuracy not yet measured
The Introduction (page unavailable) says the standard response to conflicting engineering requirements is to weigh values and make trade-offs, while technical solutions are sometimes overlooked. “Moral Residues as Motors for Technological Innovation” argues that engineers should seek more advanced technology when current technology cannot meet their obligations. The paper does not specifically say that an old-style division of labour fails to incentivise engineers to do so. (LLM, conf=0.82)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“By using Mecacci and Santoni de Sio's (2020) approach, we can more easily identify other actors whose higher-level plans and goals played or could have played a control role in the story, and may therefore have some form of responsibility in it.”
What the sources say
- Supportedcites Mecacci and Santoni de Sio's (2020) source
Split into 2 atoms — 2 confirmed, 0 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.
hedgedstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The general concern is that AI may make individual persons less able to understand, explain, and reflect upon their own and other agents' behaviour. Let us call this the moral accountability gap .”
hedgedstates a possible cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
39 more claims
Create a free account to see all 98 claims and to verify the 3 references still open.
- Probably not a substantial claim. 37
- It may state two separate claims as one. 25
- The quote only partly states this claim. 1
Citations
Cites
Street-Level Bureaucracy: Dilemmas of the Individual in Public Services
Lipsky, M. (1981) · Michigan Law Review · doi:10.2307/1288305 · Cited by 8,981
The ethics of algorithms: Mapping the debate
Mittelstadt, B. D., Allo, P., Taddeo, M., Wachter, S., & Floridi, L. (2016) · Big Data & Society · doi:10.1177/2053951716679679 · Cited by 4,567 · cited 4×
“Risks of gaps have been identified in relation not only to the learning capacities of AI but first and foremost to the opacity, complexity, and unpredictability that these systems generally display (Mittelstadt et al., 2016).”
BackgroundInfluentialThe Black Box Society: The Secret Algorithms That Control Money and Information
Pasquale, F. (2015) · doi:10.4159/harvard.9780674736061 · Cited by 3,796
Automating Inequality: How High-Tech Tools Profile, Police, and Punish the Poor
Lebovits, H. (2019) · Public Integrity · doi:10.1080/10999922.2018.1511671 · Cited by 3,279 · cited 1×
“For example, in her book Automating Inequality , Virginia Eubanks (2018) tells the stories of private contractors not being able and/or willing of disclosing the reasons for the failures/mistakes in their digital systems used by some US states for welfare benefit allocation procedures.”
BackgroundStrawson, P. F. (1962) · PhilPapers (PhilPapers Foundation) · Cited by 3,223
Developing a framework for responsible innovation
Stilgoe, J., Owen, R., & Macnaghten, P. (2013) · Research Policy · doi:10.1016/j.respol.2013.05.008 · Cited by 3,184 · cited 2×
“The responsible innovation approach invites stakeholders to be responsive to one another and to the relevant values in the process of technological development (Stilgoe et al., 2013).”
BackgroundAnalysing and Assessing Accountability: A Conceptual Framework 1
Bovens, M. (2007) · European Law Journal · doi:10.1111/j.1468-0386.2007.00378.x · Cited by 2,667 · cited 2×
“Public accountability, 5 in Mark Bovens’ definition, is a “relationship between an actor and a forum, in which the actor has an obligation to explain and to justify his or her conduct, the forum can pose questions and pass judgement, and the actor may face conse-quences” (Bovens, 2007).”
BackgroundResponsibility and Control: A Theory of Moral Responsibility
Fischer, J. M., & Ravizza, M. (1998) · doi:10.1017/cbo9780511814594 · Cited by 2,249
Can we open the black box of AI?
Castelvecchi, D. (2016) · Nature · doi:10.1038/538020a · Cited by 1,773 · cited 2×
“The way knowledge is represented in the machine and the exact way the machine distinguishes between a positive and negative diagnosis are not only inaccessible to the doctor who uses the sys-tem but also, to an important extent, to those who designed it (Castelvecchi, 2016).”
BackgroundPereboom, D. (2001) · Cambridge University Press eBooks · doi:10.1017/cbo9780511498824 · Cited by 1,519
Research works printed but not linked: 1
- Zouridis, S., Bovens, M., & Van Eck, M. (2019). Digital discretion. In T. Evans & P. Hupe (Eds.), Discretion and the quest for controlled freedom. Palgrave/MacMillan.
Cited by
Ali, S., Abuhmed, T., El-Sappagh, S., Muhammad, K., Alonso-Moral, J. M., Confalonieri, R., Guidotti, R., Del Ser, J., Díaz-Rodríguez, N., & Herrera, F. (2023) · Information Fusion · doi:10.1016/j.inffus.2023.101805 · Cited by 1,762
Agentic AI: Autonomous intelligence for complex goals–a comprehensive survey
Acharya, D. B., Kuppan, K., & Divya, B. (2025) · IEEE Access · doi:10.1109/access.2025.3532853 · Cited by 717
“Some approaches suggest an AI responsibility chain , which separates the roles and responsibilities of each stakeholder during the development and deployment process to prevent over-reliance on one individual [97].”
Esmaeilzadeh, P. (2024) · Artificial Intelligence in Medicine · doi:10.1016/j.artmed.2024.102861 · Cited by 377
REVIEWING THE ETHICAL IMPLICATIONS OF AI IN DECISION MAKING PROCESSES
Osasona, F., Amoo, O. O., Atadoga, A., Abrahams, T. O., Farayola, O. A., & Ayinla, B. S. (2024) · International Journal of Management & Entrepreneurship Research · doi:10.51594/ijmer.v6i2.773 · Cited by 188
Bleher, H., & Braun, M. (2022) · AI and Ethics · doi:10.1007/s43681-022-00135-x · Cited by 164
“Various voices in current debates on responsibility in the context of AI systems have identified a “responsibility gap”—or a multiplicity of different gaps, as Santoni de Sio and Mecacci [24] propose—as a major problem with the systems’ use [25–30].”
Background
Metadata
- Authors
- Santoni de Sio, F., & Mecacci, G.
- Affiliations
- Radboud University Nijmegen (NL)
- Venue
- Philosophy & Technology
- Published
- 2021-05-14
- Volume
- 34
- Issue
- 4
- Pages
- 1057-1084
- ISSN
- 2210-5433 (ISSN-L)
- 2210-5441
- OpenAlex
- W3160005967
- Cited by
- 490
- References
- 89
- Citations by year
- 2026: 165 · 2025: 117 · 2024: 88 · 2023: 72 · 2022: 35 · 2021: 13
- FWCI
- 64.99
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- creativecommons.org/licenses/by/4.0 (tdm)
- creativecommons.org/licenses/by/4.0 (vor)
- Locations
- Philosophy & Technology (publishedVersion, cc-by)
- Radboud Repository (Radboud University) (submittedVersion)
- Research Repository (Delft University of Technology) (submittedVersion)
- Radboud Repository (Radboud University) (submittedVersion, other-oa)
- Data Archiving and Networked Services (DANS) (submittedVersion, other-oa)
- Funders
- Nederlandse Organisatie voor Wetenschappelijk Onderzoek (MVI.16.044)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Social Sciences › Law › Digitalization, Law, and Regulation
- Social Sciences › Law › Literary and Philosophical Studies
- Keywords
- responsibility gap, artificial intelligence, moral culpability, accountability, learning automata, fatalism, deflationism, solutionism, meaningful human control, socio-technical systems, ethics of technology, legal theory, ethics
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol and qwen3.8-omni-flash).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Four Responsibility Gaps with Artificial Intelligence: Why they Matter and How to Address them. https://app.cochat.ai/papers/10.1007/s13347-021-00450-x
- The paper
- Santoni de Sio, F., & Mecacci, G. (2021). Four Responsibility Gaps with Artificial Intelligence: Why they Matter and How to Address them. Philosophy & Technology. https://doi.org/10.1007/s13347-021-00450-x
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.
