A Unified Framework of Five Principles for AI in Society
Open accessThe DOI matches its Crossref record; 24 of the 24 references it cites were found in a registry or an index; 1 of the 6 sources its claims cite were checked for support.
Abstract
Artificial Intelligence (AI) is already having a major impact on society. As a result, many organizations have launched a wide range of initiatives to establish ethical principles for the adoption of socially beneficial AI. Unfortunately, the sheer volume of proposed principles threatens to overwhelm and confuse. How might this problem of ‘principle proliferation’ be solved? In this paper, we report the results of a fine-grained analysis of several of the highest-profile sets of ethical principles for AI. We assess whether these principles converge upon a set of agreed-upon principles, or diverge, with significant disagreement over what constitutes ‘ethical AI.’ Our analysis finds a high degree of overlap among the sets of principles we analyze. We then identify an overarching framework consisting of five core principles for ethical AI. Four of them are core principles commonly used in bioethics: beneficence, non-maleficence, autonomy, and justice. On the basis of our comparative analysis, we argue that a new principle is needed in addition: explicability, understood as incorporating both the epistemological sense of intelligibility (as an answer to the question ‘how does it work?’) and in the ethical sense of accountability (as an answer to the question: ‘who is responsible for the way it works?’). In the ensuing discussion, we note the limitations and assess the implications of this ethical framework for future efforts to create laws, rules, technical standards, and best practices for ethical AI in a wide range of contexts.
Overview
CoChat Summary
TL;DR
The paper analyzes six high-profile AI ethics documents containing 47 principles and finds a high degree of overlap, synthesizing a unified framework of five principles: beneficence, non-maleficence, autonomy, justice, and explicability. It argues that explicability—incorporating intelligibility and accountability—is the missing fifth principle needed to complement the four bioethics principles.
Key findings
- A comparative analysis of six initiatives (Asilomar, Montreal, IEEE, EGE, AIUK, Partnership on AI) yielded 47 principles; the five-principle framework captures every one of the 47 principles.
- Four bioethics principles—beneficence, non-maleficence, autonomy, and justice—adapt well to AI ethics but are contested and not exhaustive; a fifth principle, explicability, is needed, combining intelligibility (“how does it work?”) and accountability (“who is responsible for the way it works?”).
- Each of the five principles is included in almost every statement of principles analyzed (Table 1); the framework was later largely adopted by the European Commission’s HLEGAI and influenced the OECD Recommendation reaching 42 countries.
- The paper finds convergence impressive and reassuring but cautions that the underlying meaning of each principle is contested, with similar terms often used to mean different things.
Methods
- Comparative, fine-grained analysis of six high-profile AI ethics documents selected as recent (published within the last three years), directly relevant to AI and its impact on society, and highly reputable, multistakeholder organizations with at least national scope.
- The documents analyzed are the Asilomar AI Principles, Montreal Declaration, IEEE Ethically Aligned Design v2 General Principles, EGE Statement, UK House of Lords AI Committee report, and Partnership on AI Tenets; together they yield 47 principles.
- The analysis compares these principles against the four core bioethics principles and identifies overlap and convergence, then synthesizes an overarching five-principle framework.
Limitations
- The paper states that each of the six sets of principles emerged either from initiatives with global scope or from within western liberal democracies; for broader applicability, the framework would benefit from perspectives of regions and cultures presently un- or under-represented, notably China.
- The paper states that the detected convergence demands caution because the underlying meaning of each principle is contested, with similar terms often used to mean different things.
- Evident from the text: the authors disclose potential conflicts—Floridi chaired the AI 4 People project and is a member of the European Commission’s HLEGAI, while Cowls was rapporteur—and funding from Facebook and Google, which may bear on interpretation.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures
Verification
Claims the paper makes
1 Valid · 1 Not valid · 5 Not checked · 17 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Indeed, the framework played a valuable role in the work of AI 4 People, Europe's first global forum on the social impact of AI, which recently adopted it to propose 20 concrete recommendations for a 'Good AI Society' to the European Commission (Floridi et al., 2018 ).”
What the sources say
- Partly supportedcites Floridi et al., 2018 sourceExperimental — accuracy not yet measured
The Introduction and Conclusion say AI4People used its ethical principles to formulate 20 recommendations for a Good AI Society. The Recommendations section says some are for national or European policymakers, but our copy does not say the recommendations were proposed to the European Commission specifically. (LLM, conf=0.86)
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
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Moreover, each principle is included in almost every statement of principles we analyzed (see Table 1 below).”
plain- 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.
In the paper
“Since then it has been largely adopted by the Ethics Guidelines for Trustworthy AI published by the European Commission's High-Level Expert Group on Artificial Intelligence (HLEGAI 2018 and 2019 ), which in turn has influenced the OECD's Recommendation of the Council on Artificial Intelligence (OECD 2019 ), reaching 42 countries 6 (see Table 1 ).”
What the sources say
- Not matchedcites OECD 2019
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
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.
- Not matched
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“It is not therefore surprising that the principle of autonomy is explicitly stated in four of the six documents.”
plain- 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.
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“It is important to note that each of the six sets of ethical principles for AI that we analyzed emerged either from initiatives with global scope, or from within western liberal democracies.”
plain- 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.
9 more claims
Create a free account to see all 24 claims and every source verdict.
- It may state two separate claims as one. 3
- Probably not a substantial claim. 7
- The quote only partly states this claim. 1
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Conversely, HAKI maintains that authorship remains exclusively human, as AI lacks intentionality, critical reasoning, and epistemic responsibility [34].”
From Developing the HAKI model as a conceptual framework for AI-assisted academic research (2026) · cites this paper as [34]
Experimental — accuracy not yet measuredThe check wasn’t confident either way, so it gives no verdict — never disputed.
In the citing paper
“Conversely, HAKI maintains that authorship remains exclusively human, as AI lacks intentionality, critical reasoning, and epistemic responsibility [34].”
From Developing the HAKI model as a conceptual framework for AI-assisted academic research (2026) · cites this paper as [34]
Experimental — accuracy not yet measuredIn the citing paper
“While these guidance documents vary in language (Jobin et al., 2019), they converge on a limited set of principles (Floridi & Cowls, 2019).”
From Auditing of AI: Legal, Ethical and Technical Approaches (2023) · cites this paper as Floridi & Cowls, 2019
Citations
Cites
Principles of Biomedical Ethics
(2012) · Thieves of Virtue · doi:10.7551/mitpress/9079.003.0009 · Cited by 13,453
I.—COMPUTING MACHINERY AND INTELLIGENCE
TURING, A. M. (1950) · Mind · doi:10.1093/mind/lix.236.433 · Cited by 10,290
Computing Machinery and Intelligence (1950)
Turing, A. (2004) · The Essential Turing · doi:10.1093/oso/9780198250791.003.0017 · Cited by 6,118
Floridi, L., Cowls, J., Beltrametti, M., Chatila, R., Chazerand, P., Dignum, V., Luetge, C., Madelin, R., Pagallo, U., Rossi, F., et al. (2018) · Minds and Machines · doi:10.1007/s11023-018-9482-5 · Cited by 4,774
Self-citationThe Ethics of AI Ethics: An Evaluation of Guidelines
Hagendorff, T. (2020) · Minds and Machines · doi:10.1007/s11023-020-09517-8 · Cited by 1,911
The Ethics of AI Ethics -- An Evaluation of Guidelines
Hagendorff, T. (2019) · arXiv · doi:10.48550/arxiv.1903.03425 · Cited by 1,907
The grand challenges of Science Robotics
Yang, G.-Z., Bellingham, J., Dupont, P. E., Fischer, P., Floridi, L., Full, R., Jacobstein, N., Kumar, V., McNutt, M., Merrifield, R., et al. (2018) · Science Robotics · doi:10.1126/scirobotics.aar7650 · Cited by 1,177
Self-citationHow AI can be a force for good
Taddeo, M., & Floridi, L. (2018) · Science · doi:10.1126/science.aat5991 · Cited by 760
Self-citationFloridi, L. (2013) · Oxford University Press eBooks · doi:10.1093/acprof:oso/9780199641321.001.0001 · Cited by 651
Self-citationTranslating Principles into Practices of Digital Ethics: Five Risks of Being Unethical
Floridi, L. (2019) · Philosophy & Technology · doi:10.1007/s13347-019-00354-x · Cited by 524
Self-citation
Research works printed but not linked: 2
- The IEEE Initiative on Ethics of Autonomous and Intelligent Systems ( 2017 ). Ethically Aligned Design, v 2 . https://ethicsinaction.ieee.org
- McCarthy, J., Minsky, M. L., Rochester, N. and Shannon, C. E.. ( 2006 ). A proposal for the Dartmouth summer research project on artificial intelligence, August 31 , 1955 . AI Magazine 27 ( 4 ): 12 .
Cited by
Morley, J., Floridi, L., Kinsey, L., & Elhalal, A. (2020) · Science and Engineering Ethics · doi:10.1007/s11948-019-00165-5 · Cited by 882
“It is, therefore, a necessary building block in the creation of an environment that fosters ethical, responsible, and beneficial ML, especially as it also indicates the possibility of a time when the distractive risk of ethics shopping 13 (Floridi 2019c) will be lessened.”
BackgroundMethodInfluentialSelf-citationTrustworthy Artificial Intelligence: A Review
Kaur, D., Uslu, S., Rittichier, K. J., & Durresi, A. (2023) · ACM Computing Surveys · doi:10.1145/3491209 · Cited by 709
“Over the past few years, research institutes, private organizations, and government agencies have proposed various guidelines and frameworks to make AI trustworthy [60, 68, 69, 154, 198, 224].”
BackgroundTrustworthy AI: From Principles to Practices
Li, B., Qi, P., Liu, B., Di, S., Liu, J., Pei, J., Yi, J., & Zhou, B. (2023) · ACM Computing Surveys · doi:10.1145/3555803 · Cited by 612
“A number of enterprises [57, 62, 136, 166, 254, 338], academia [122, 199, 218, 301, 322], public sectors, and organizations [9, 210, 334] have recently identified these factors and summarized them as principles of AI trustworthiness.”
BackgroundTrustworthy artificial intelligence
Thiebes, S., Lins, S., & Sunyaev, A. (2021) · Electronic Markets · doi:10.1007/s12525-020-00441-4 · Cited by 595
“Toward this end, “ [one] must be able to understand the good or harm [AI] is actually doing to society, and in which ways ” (Floridi and Cowls 2019) for it to be beneficent and non-maleficent.”
BackgroundInfluentialWach, K., Duong, C. D., Ejdys, J., Kazlauskaitė, R., Korzynski, P., Mazurek, G., Paliszkiewicz, J., & Ziemba, E. (2023) · Entrepreneurial Business and Economics Review · doi:10.15678/eber.2023.110201 · Cited by 584
“While ChatGPT and GAI have the potential to revolutionize how we approach data analysis and report generation, they also raise significant doubts and questions about their potential effects on society, including ethics, privacy, and security (Floridi & Cowls, 2019; Florek-Paszkowska et al ., 2021).”
Background
Metadata
- Authors
- Floridi, L., & Cowls, J.
- Affiliations
- University of Oxford (GB)
- The Alan Turing Institute (GB)
- Venue
- Harvard Data Science Review
- Publisher
- MIT Press
- Published
- 2019-06-23
- Volume
- 1
- Issue
- 1
- ISSN
- 2644-2353 (ISSN-L)
- OpenAlex
- W2954266614
- Cited by
- 1,631
- References
- 22
- Citations by year
- 2026: 600 · 2025: 463 · 2024: 207 · 2023: 127 · 2022: 98 · 2021: 80 · 2020: 52 · 2019: 3
- FWCI
- 101.67
- Type
- journal article
- Language
- en
- Open access
- diamond
- doi.org
- Licence
- cc-by
- Locations
- Harvard Data Science Review (publishedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
- Social Sciences › Law › Digitalization, Law, and Regulation
- Keywords
- AI ethics, beneficence, non-maleficence, autonomy, justice, explainability, accountability, intelligibility, bioethics, societal impact of AI
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).
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: A Unified Framework of Five Principles for AI in Society. https://app.cochat.ai/papers/10.1162/99608f92.8cd550d1
- The paper
- Floridi, L., & Cowls, J. (2019). A Unified Framework of Five Principles for AI in Society. Harvard Data Science Review. https://doi.org/10.1162/99608f92.8cd550d1
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.

