Diversity, Equity, and Inclusion in Artificial Intelligence: An Evaluation of Guidelines
Open accessThe DOI matches its Crossref record; 59 of the 59 references it cites were found in a registry or an index; none of the 81 sources its claims cite is checked for support yet.
Abstract
Artificial intelligence (AI) is present everywhere in the lives of individuals. Unfortunately, several cases of discrimination by AI systems have already been reported. Scholars have warned on risks of AI reproducing existing inequalities or even amplifying them. To tackle these risks and promote responsible AI, many ethics guidelines for AI have emerged recently, including diversity, equity, and inclusion (DEI) principles and practices. However, little is known about the DEI content of these guidelines, and to what extent they meet the most relevant accumulated knowledge from DEI literature. We performed a semi-systematic literature review of the AI guidelines regarding DEI stakes and analyzed 46 guidelines published from 2015 to today. We fleshed out the 14 DEI principles and the 18 DEI practices recommended underlying these 46 guidelines. We found that the guidelines mostly encourage one of the DEI management paradigms, namely fairness, justice, and nondiscrimination, in a limited compliance approach. We found that narrow technical practices are favored over holistic ones. Finally, we conclude that recommended practices for implementing DEI principles in AI should include actions aimed at directly influencing AI actors’ behaviors and awareness of DEI risks, rather than just stating intentions and programs.
Overview
CoChat Summary
TL;DR
This semi-systematic review analyzes 46 AI ethics guidelines published from 2015 onward and finds that their DEI content is dominated by fairness, justice, and nondiscrimination in a limited compliance approach. It concludes that guidelines favor narrow technical practices over holistic organizational ones and rarely target AI actors’ day-to-day behaviors or awareness of DEI risks.
Key findings
- The review identified 14 DEI principles and 18 DEI practices across 46 guidelines. Of the 14 principles, 12 (86%) map to the “fair representation and treatment and nondiscrimination” paradigm, while only 2 (14%) reflect a “valuing diversity” or synergy perspective.
- The most cited DEI principles were Equity/Fairness (76%, 35 sources), Nondiscrimination (57%, 26), and Diversity in learning datasets/Representativeness (50%, 23). The most targeted groups were female (N = 34), disabled or health vulnerable people (N = 34), and racial or ethnic minorities (N = 32).
- In the AIS development process, the most recommended practices were verification and validation of AIS results (63%, 29), AIS learning sources and dataset (57%, 26), transparency of AIS decisions (54%, 25), and explicability of AIS decisions (52%, 24).
- For AIS design organizations, accountability/responsibility for AIS was recommended by 57% (26), team diversity in AIS companies by 50% (23), and DEI training/awareness by 43% in Table 5 (20; described as 50% in the text). Only one of eight organizational practices directly addressed the praxis dimension—objectives/incentives for compliance with DEI principles (13%, 6)—and no guideline included penalties for noncompliance.
Methods
- Semi-systematic literature review adapted from PRISMA, using a two-step search: prior comprehensive AI ethics reviews plus keyword-based web search. Inclusion criteria were documents published in 2015 or later, written in English, and from reliable influential sources such as governmental/intergovernmental reports, professional AI associations, and influential private AI companies; the final sample was 46 documents.
- Content extraction and categorization used deductive categories based on DEI expertise, followed by inductive identification of additional categories. This yielded 14 DEI principles and 18 DEI practices, with 361 extracted entries for principles and 588 for practical recommendations.
- Two research graduate students coded independently; agreement was almost perfect for principles (k = 0.85, p < .05) and AIS development process (k = 0.84, p < .05), and substantial for AIS design organizations (k = 0.80, p < .05). A third researcher specializing in DEI management in AI inspected the results.
Limitations
- Paper-stated: the review selected only English-language sources, which may explain the strong Western domination in the results; the authors also note that no document, code, or charter dedicated specifically to DEI in AI was found within their criteria.
- Paper-stated: the authors call for further research on whether intentions and pro-DEI programs alone can influence AI actors’ DEI behaviors, or whether binding regulation is needed; they also note the lack of diversity among guideline authors (60.8% male, 74.2% white, 78.3% nationals of Western countries) and within the AI community.
- Evident from the text: the analysis is based on manual coding by two graduate students with a third researcher inspecting results, and the sample is limited to 46 English-language guidelines, so the categorization and geographic coverage may be shaped by those constraints.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures
Verification
Claims the paper makes
0 Valid · 1 Not valid · 71 Not checked · 119 Own results and interpretations
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
“DEI principles in AI always target certain groups to benefit from the erected principles, such as social justice or fairness.”
§ Inclusion in Artificial Intelligence: An Evaluation of Guidelines > Targeted Groups in DEI Principles in AI
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.
In the paper
“Results show an almost perfect agreement for the principles (k = 0.85; p < .05)and the AIS development process (k = 0.84; p < .05),and a substantial agreement for the regarding AIS design organizations (k = 0.80; p < .05)(Landis and Koch 1977).”
§ Inclusion in Artificial Intelligence: An Evaluation of Guidelines > Content Extraction and Categorization
What the sources say
- Not checked yetcites Landis and Koch 1977
plainstates an association- 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 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
“Practices for ensuring more diverse and representative”
§ Inclusion in Artificial Intelligence: An Evaluation of Guidelines > The AIS Development Process
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.
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
“The respect of human dignity and social justice (referred, respectively, to by 35% and 33% of guidelines) are also pointed out, referring to a consideration of all individuals without distinction.”
§ Inclusion in Artificial Intelligence: An Evaluation of Guidelines > DEI Principles in AI Ethical Guidelines
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 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
“Also, 26% percent of guidelines consider the importance of addressing the digital divide, which entails that all have access to AISs developed, with adequate literacy and understanding, fighting against barriers to access and empowering beneficiaries.”
§ Inclusion in Artificial Intelligence: An Evaluation of Guidelines > DEI Principles in AI Ethical Guidelines
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.
79 more claims
Create a free account to see all 191 claims and to verify the 81 references still open.
- Probably not a substantial claim. 59
- The quote only partly states this claim. 13
- It may state two separate claims as one. 65
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 cited source says what the claim attributes to it.
In the citing paper
“Incorporating DEI principles must be perceived as an essential priority, particularly given a lack of representation in current regulatory guidelines for AI applications50,83.”
From Bias recognition and mitigation strategies in artificial intelligence healthcare applications (2025) · cites this paper as [50]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Inclusion in Artificial Intelligence: An Evaluation of Guidelines > The Rather Technical Operationalization of DEI Principles in AI Practices (R2a)
“7 out of 10 practices devoted to the AIS development process are recommended by 40% and more of guidelines”
§ Inclusion in Artificial Intelligence: An Evaluation of Guidelines > Conclusion
“| Functional design of AIS | 22 | 48% | | Technical design of AIS | 18 | 39% | | Correction of biased AIS results | 18 | 39% | | Mechanisms of redress for users/beneficiaries | 16 | 35% |”
Citations
Cites
The Measurement of Observer Agreement for Categorical Data
Landis, J. R., & Koch, G. G. (1977) · Biometrics · doi:10.2307/2529310 · Cited by 81,148
Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement
Moher, D., Liberati, A., Tetzlaff, J., Altman, D. G., & The PRISMA Group (2009) · PLoS Medicine · doi:10.1371/journal.pmed.1000097 · Cited by 64,042
Tranfield, D., Denyer, D., & Smart, P. (2003) · British Journal of Management · doi:10.1111/1467-8551.00375 · Cited by 15,386
Literature review as a research methodology: An overview and guidelines
Snyder, H. (2019) · Journal of Business Research · doi:10.1016/j.jbusres.2019.07.039 · Cited by 11,524
The global landscape of AI ethics guidelines
Jobin, A., Ienca, M., & Vayena, E. (2019) · Nature Machine Intelligence · doi:10.1038/s42256-019-0088-2 · Cited by 6,220
Mitchell, R. K., Agle, B. R., & Wood, D. J. (1997) · The Academy of Management Review · doi:10.2307/259247 · Cited by 5,039
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
What is AI Literacy? Competencies and Design Considerations
Long, D., & Magerko, B. (2020) · Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3313831.3376727 · Cited by 3,241 · cited 1×
“If AI literacy begins to be a subject of research and recommendations (Long and Magerko 2020; Ng et al. 2021), such research and recommendations should also explore the technological distance and literacy from diverse economic and cultural contexts.”
BackgroundUnderstanding HRM–Firm Performance Linkages: The Role of the “Strength” of the HRM System
Bowen, D. E., & Ostroff, C. (2004) · Academy of Management Review · doi:10.5465/amr.2004.12736076 · Cited by 2,996 · cited 1×
“When widely perceived and interpreted by organizational incumbents, these messages create a strong climate (Bowen and Ostroff 2004), that would lead them to adopt the desired behaviors toward DEI.”
Background(2019) · Cited by 2,637
Research works printed but not linked: 44
- Accenture. 2019. Responsible AI and robotics: An ethical framework. Accenture, UK.
- ACM. 2017. Statement on algorithmic transparency and accountability. Washington, DC: Association for Computing Machinery US Public Policy Council.
- Avila, R., A. Brandusescu, J. O. Freuler, and D. Takur. 2018. Artifcial intelligence: Open questions about gender inclusion. Argentina: World Wide Web Foundation.
- Beijing Academy of Artificial Intelligence. 2019. Beijing AI Principles. Beijing Academy of Artificial Intelligence (BAAI), Beijing: Beijing Academy of Artifcial Intelligence.
- Benessaieh, K. 2022. Le Québec se classe 7e au monde. La Presse, 29(2): March 9th.
- Dawson, D., E. Schleiger, J. Horton, J. McLaughlin, C. Robinson, G. Quezada, J. Scowcroft, and S. Hajkowicz. 2019. Artificial intelligence: Australia’s ethics framework. Data61 CSIRO, Australia.
- DeepMind. 2019. Ethics and society principles. London: DeepMind.
- Deloitte. 2020. Trustworthy AI. USA: The Deloitte AI Institute.
- Diakopoulos, N., Friedler S, Arenas M, Barocas S, Hay M, Howe B, Jagadish HV, Unsworth K, Sahuguet A, Venkatasubramanian S,et al. 2019. Principles for accountable algorithms and a social impact statement for algorithms. Fairness, Accountability, and Transparency in Machine Learning (FAT/ML). https://www.fatml.org/resources/principles-for-accountable-algorithms
- Dutch Artificial Intelligence. 2018. Dutch artificial intelligence manifesto. The Netherlands: Special Interest Group on Artificial Intelligence.
- Eubanks, V. 2018. Automating inequality: How high-tech tools profle, police, and punish the poor. New York: St. Marting’s Press.
- European Commission. 2020. The Assessment List for Trustworthy Artificial Intelligence for self assessment. In Directorate-General for Communications Networks. Content and Technology, Publications Office.
- Freeman, R. E. 1984. Strategic management: A stakeholder theory. Journal of Management Studies 39 (1):1–21.
- Future of Life Institute. 2017. Asilomar AI principles. California: Future of Life Institute.
- The Future Society. 2018. The future society, law & society initiative, principles for the governance of AI. Policy Research, The Law & Society Initiative.
- Google. 2018. Artificial intelligence at google: Our principles. Google AI.
- Government of the Republic of Korea. 2017. Mid- to long-term master plan in preparation for the intelligent information society: Managing the fourth industrial revolution. Korea: Government of the Republic of Korea Interdepartmental Exercise.
- GOV.UK. 2019. Initial code of conduct for data-driven health and care technology. United Kingdom: GOV.UK.
- Green Digital Working Group. 2016. Position on robotics and artificial intelligence. The Green Digital Working Group, Europe.
- IBM. 2018. Everyday ethics for artificial intelligence, IBM design for AI. USA.
- IEEE. 2016. Ethically aligned design: a vision for prioritizing human well-being with autonomous and intelligent systems, version 1. IEEE Advancing Technology for Humanity. https://standards.ieee.org/wp-content/uploads/import/documents/other/ead_v1.pdf
- IEEE. 2019. Ethically aligned design: A vision for prioritizing human well-being with autonomous and intelligent systems, First Edition ed, IEEE Advancing Technology for Humanity. https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9398613
- Information Accountability Foundation. 2015. Unified ethical frame for big data analysis: IAF big data ethics initiative. Part A.
- Institute of Business Ethics. 2018. Business ethics and artificial intelligence. London: Institute of Business Ethics.
- ITI. 2017. AI policy principles. Information Technology Industry Council. https://www.itic.org/public-policy/ITIAIPolicyPrinciplesFINAL.pdf
- Japanese Society for Artificial Intelligence. 2017. Ethical Guidelines. In The Japanese Society for Artificial Intelligence, Japan.
- Klarsfeld, A., L. A. E. Booysen, E. Ng, I. Roper, and A. Tatli. 2014. Perspectives from 16 countries on diversity and equal treatment at work: An overview and transverse questions country perspectives on diversity and equal treatment. Cheltenham: Edward Elgar Publishing.
- Leaders of the G7. 2018. Common vision for the future of artificial intelligence, G7. Charlevoix.
- Microsoft Corporation. 2019. Microsoft AI principles. USA: Microsoft Corporation.
- Ministry of State for Science and Technology Policy. 2017. Report on Artificial Intelligence and Human Society: Unofficial Translation. In Advisory Board on Artificial Intelligence and Human Society, Japan.
- Montréal Declaration. 2018. Montréal declaration for a responsible development of artificial intelligence. Montreal. https://www.montrealdeclaration-responsibleai.com/
- Partnership on AI. 2018. Advancing positive outcomes for people and society.
- PwC. 2019. The responsible AI framework. PwC.
- Royal Society. 2017. Machine learning: The power and promise of computers that learn by example. London: The Royal Society.
- Thomas, D. A., and R. J. Ely. 1996. Making differences matter: A new paradigm for diversity management. Harvard Business Review 74 (5):79–90.
- Toronto declaration. 2018. The toronto declaration, protecting the right to equality and non-discrimination in machine learning systems. Toronto: Human Rights Watch.
- UNESCO. 2021. Recommendation on the ethics of artificial intelligence. Paris: SHS/BIO/REC-AIETHICS/2021.
- UNI Global Union. 2017. Top 10 Principles for Ethical AI. Switzerland: UNI Global Union.
- van Est, R., and J. Gerritsen. 2017. Human rights in the robot age: Challenges arising from the use of robotics, AI, and virtual and augmented reality. The Netherlands: Rathenau Instituut.
- Villani, C. 2018. For a meaningful artificial intelligence: Towards a French and European strategy. France: AI for Humanity.
- Wallach, W., and G. E. Marchant 2018. An agile ethical/legal model for the international and national governance of AI and robotics. Aies Conference on Artificial Intelligence, Ethics and Society, New Orleans, USA.
- West, S. M., M. Whittaker, and K. Crawford. 2019. Discriminating systems: Gender, race and power in AI. AI Now Institute.
- Whittaker, M., Crawford K, Dobbe R, Fried G, Kaziunas E, Mathur V, West SM, Richardson R, Schultz J, Schwartz O et al. 2018. AI now report 2018. New York: AI Now Institute.
- World Economic Forum. 2018. How to Prevent Discriminatory Outcomes in Machine Learning. In World Economic Forum Global Future Council on Human Rights 2016-18, REF, 120318–case 00040065 . Switzerland.
Cited by
Bias recognition and mitigation strategies in artificial intelligence healthcare applications
Hasanzadeh, F., Josephson, C. B., Waters, G., Adedinsewo, D., Azizi, Z., & White, J. A. (2025) · npj Digital Medicine · doi:10.1038/s41746-025-01503-7 · Cited by 320
AI and the quest for diversity and inclusion: a systematic literature review
Shams, R. A., Zowghi, D., & Bano, M. (2025) · AI and Ethics · doi:10.1007/s43681-023-00362-w · Cited by 192
Onciul, R., Tataru, C.-I., Dumitru, A. V., Crivoi, C., Serban, M., Covache-Busuioc, R.-A., Radoi, M. P., & Toader, C. (2025) · Journal of Clinical Medicine · doi:10.3390/jcm14020550 · Cited by 113
“Ensuring that future AI systems are inclusive and equi-table will require concerted efforts to collect diverse, representative data [299].”
Marko, J. G. O., Neagu, D. C., & Anand, P. B. (2025) · BMC Medical Informatics and Decision Making · doi:10.1186/s12911-025-02884-1 · Cited by 153
Viberg, O., Kizilcec, R. F., Wise, A. F., Jivet, I., & Nixon, N. (2024) · British Journal of Educational Technology · doi:10.1111/bjet.13507 · Cited by 98
Metadata
- Authors
- Cachat-Rosset, G., & Klarsfeld, A.
- Affiliations
- Department of Management, Faculty of Business Administration, Université Laval, Québec, QC, Canada
- Head of Laboratory Work, employment and Health, Toulouse Business School Toulouse, Toulouse, France
- Venue
- Applied Artificial Intelligence
- Publisher
- Informa UK Limited
- Published
- 2023-02-22
- Volume
- 37
- Issue
- 1
- ISSN
- 0883-9514 (ISSN-L)
- 1087-6545
- OpenAlex
- W4321494006
- Semantic Scholar
- b690c7900b0e9c0295cdcee7491824dfca128c50
- Cited by
- 138
- References
- 58
- Citations by year
- 2026: 32 · 2025: 64 · 2024: 33 · 2023: 9
- FWCI
- 34.61
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Applied Artificial Intelligence (publishedVersion, cc-by-nc)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- 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 › Business, Management and Accounting › Organizational Behavior and Human Resource Management › AI and HR Technologies
- Keywords
- diversity, equity, and inclusion, algorithmic discrimination, responsible AI, compliance approach
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.
A language model drew the claims from the paper on 2026-10-05, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-05 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
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: Diversity, Equity, and Inclusion in Artificial Intelligence: An Evaluation of Guidelines. https://app.cochat.ai/papers/10.1080/08839514.2023.2176618
- The paper
- Cachat-Rosset, G., & Klarsfeld, A. (2023). Diversity, Equity, and Inclusion in Artificial Intelligence: An Evaluation of Guidelines. Applied Artificial Intelligence. https://doi.org/10.1080/08839514.2023.2176618
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.



