CoChat
Sign Up
Cited by 331Ethics and Social Impacts of AIDigitalization, Law, and RegulationExplainable Artificial Intelligence (XAI)

Artificial intelligence governance: Ethical considerations and implications for social responsibility

Open access
Camilleri, Mark Anthony·2024·Expert Systems
DOI 10.1111/exsy.13406
Open this paper in CoChat — free accountOpen PDF

Abstract

A number of articles are increasingly raising awareness on the different uses of artificial intelligence (AI) technologies for customers and businesses. Many authors discuss about their benefits and possible challenges. However, for the time being, there is still limited research focused on AI principles and regulatory guidelines for the developers of expert systems like machine learning (ML) and/or deep learning (DL) technologies. This research addresses this knowledge gap in the academic literature. The objectives of this contribution are threefold: (i) It describes AI governance frameworks that were put forward by technology conglomerates, policy makers and by intergovernmental organizations, (ii) It sheds light on the extant literature on ‘AI governance’ as well as on the intersection of ‘AI’ and ‘corporate social responsibility’ (CSR), (iii) It identifies key dimensions of AI governance, and elaborates about the promotion of accountability and transparency; explainability, interpretability and reproducibility; fairness and inclusiveness; privacy and safety of end users, as well as on the prevention of risks and of cyber security issues from AI systems. This research implies that all those who are involved in the research, development and maintenance of AI systems, have social and ethical responsibilities to bear toward their consumers as well as to other stakeholders in society.

AI snapshot

Written from the paper’s full text.

TL;DR

This paper reviews AI governance frameworks and the literature on AI governance and AI–CSR, identifying accountability/transparency, explainability/interpretability/reproducibility, fairness/inclusiveness, privacy/safety, and security/robustness as key dimensions. It argues that everyone involved in AI research, development, and maintenance has social and ethical responsibilities to consumers and other stakeholders.

Key findings

  • The systematic review found 13 Scopus articles on “artificial intelligence governance,” 12 of which were also indexed in Web of Science; 75% of captured articles used secondary research methodologies (discursive contributions and/or literature reviews).
  • For the intersection of “artificial intelligence” and “corporate social responsibility,” the search returned 36 Scopus articles; 14 were excluded, leaving 22 articles, and 41% of those used secondary research methodologies.
  • The paper describes governance frameworks from the EU AI Act, Singapore’s National AI Strategy, the US AI Bill of Rights, IEEE, OECD, IBM, and Microsoft, including the EU’s four risk levels: minimal, limited, high-risk, and unacceptable.
  • It identifies key AI governance dimensions: accountability and transparency; explainability, interpretability, and reproducibility; fairness and inclusiveness; privacy and safety for consumers; and security and robustness of AI systems, including prevention of risks and cyber security issues.

Methods

  • Design: systematic literature review (SLR) of articles focused on AI governance and CSR, intended to be replicable and trustworthy.
  • Data/sample: Scopus search for TITLE-ABS-KEY (“artificial intelligence governance”), limited to articles, English, journals, and publication years 2019–2023, yielding 13 Scopus articles (12 also in WOS Core Collections). A second Scopus/WOS search for TITLE-ABS-KEY (“artificial intelligence” AND “corporate social responsibility”) with the same limits yielded 36 articles, of which 14 were excluded, leaving 22.
  • Analysis: bibliographic review of authors, objectives, methodologies, and prominent keywords, plus discursive synthesis of regulatory principles and a proposed responsible AI governance framework.

Limitations

  • Stated: most contributions on AI governance were discursive and/or literature reviews, with few primary research activities; the paper calls for future primary research using different designs, methodologies, and sampling frames.
  • Stated: there are limited contributions focused on AI governance frameworks providing substantive and reflexive guidelines, and few jurisdictions have formalized AI policies and governance frameworks.
  • Evident: the SLR is restricted to Scopus/WOS-indexed English-language journal articles from 2019–2023 using specific keyword queries; the proposed AI governance framework is conceptual and is not empirically tested in the paper.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1111/exsy.13406

Cites

85 of 85 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 87 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    Most articles (41%) featured in the SLR on CSR and AI involved secondary research methodologies.

    “Again, Table 3 confirms that most articles (41%) that were featured in this SLR exercise involved secondary research methodologies.”
    Quote in the paperNot a substantial claim
  • Findingown data

    The most popular keywords on the intersection of AI and CSR were Corporate Social Responsibility, Artificial Intelligence, Sustainability, Machine Learning, Sustainable Development, Business Ethics, Corporate Governance, Ethics, Health Care and Human Resource.

    “This bibliographic study indicates that the most popular keywords on the intersection of AI and CSR were: Corporate Social Responsibility, Artificial Intelligence, Sustainability, Machine Learning, Sustainable Development, Business Ethics, Corporate Governance, Ethics, Health Care and Human Resource.”
    Quote in the paperNot a substantial claim
  • Findingown data

    The Scopus search for articles with 'artificial intelligence governance' in title, abstract or keywords, limited to English journal articles from 2019-2023, returned 13 articles.

    “The results indicated that there were thirteen (13) articles through Scopus, that featured the specified keywords.”
    Quote in the paperSubstantial claim
  • Findingown data

    Most articles (75%) captured in the AI governance review involved secondary research methodologies (discursive contributions and/or literature reviews).

    “Table 2 clearly indicates that most articles (75%) that were captured through this review involved secondary research methodologies as they were discursive contributions and/or featured literature reviews.”
    Quote in the paperNot a substantial claim
  • Findingown data

    The top ten keywords used by researchers investigating AI governance were Artificial Intelligence, Artificial Intelligence Governance, AI Governance, AI, AI Ethics, Artificial Intelligence Ethic, Artificial Intelligence Systems, Ethical Technology, Decision Making and AI Systems.

    “The SLR reported that the most prominent (top ten) keywords that were used by researchers who investigated AI governance were Artificial Intelligence, Artificial Intelligence Governance, AI Governance, AI, AI Ethics, Artificial Intelligence Ethic, Artificial Intelligence Systems, Ethical Technology, Decision Making and AI Systems.”
    Quote in the paperNot a substantial claim
  • Classified as probably not a substantial claim. 71
  • May combine two claims. 27

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingcitation

    ChatGPT indicated that its responses are generated algorithmically and may not be completely accurate or fully address the question being asked.

    “AI language model, ChatGPT strives to provide accurate and helpful responses based on the input it received.However, it's important to note that my responses are generated algorithmically, and therefore, there may be instances where my responses are not completely accurate or may not fully address the question being asked.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingcitation

    ChatGPT indicated that its responses are based on the data available at its training and knowledge cutoff date, so new developments or information after that date may not be reflected.

    “Additionally, my responses are based on the data available to me at the time of my training and knowledge cutoff date, so any new developments or information that have emerged since then, may not be reflected in my responses.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingnone given

    When questioned about its cutoff date, ChatGPT indicated that events, developments or information after September 2021 were not reflected in its responses.

    “When it was questioned about its cutoff date, ChatGPT indicated that any events, developments or information that occurred after September 2021 was not reflected in its responses.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Twelve of the 13 Scopus articles were also indexed in Web of Science Core Collections (A&HCI, ESCI, SCIE and/or SSCI).

    “Twelve (12) of these publications were also indexed through Web of Science's (WOS) Core Collections including in Arts & Humanities Citation Index (A&HCI), Emerging Sources Citation Index (ESCI), Science Citation Index (SCIE) and/or in Social Science Citation Index (SSCI).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Table 3 features 22 articles on the intersection of artificial intelligence and CSR.

    “Table 3 features all (22) articles on the intersection of artificial intelligence and CSR.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Fourteen articles were excluded because they were not focused on AI or CSR, or were not published in a WOS Core Collection in addition to Scopus.

    “However, fourteen (14) articles were excluded as they were not focused on AI or on corporate social responsibility (CSR) paradigms.Alternatively, the discarded publications were not published in one of WOS's Core Collections (in addition to Scopus).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Findingnone given

    At the moment there are a few jurisdictions that have formalized their AI policies and governance frameworks.

    “It reported that, at the moment there are a few jurisdictions that have formalized their AI policies and governance frameworks.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingcitation

    The review reports that proponents are raising awareness of AI benefits for individuals and organizations, while scholars and stakeholders including policy makers are raising concerns about possible perils.

    “It reported that its proponents are raising awareness on the benefits of AI systems for individuals as well as for organizations.At the same time, it suggests that a number of scholars and other stakeholders including policy makers, are raising their concerns about its possible perils (e.g., Berente et al., 2021;Gonzalez et al., 2020;Zhang & Lu, 2021).”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Berente et al., 2021

      Not checked in the free run.

    • cites Gonzalez et al., 2020

      Not checked in the free run.

    • cites Zhang & Lu, 2021

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingnone given

    The review indicates that more academics and practitioners are increasingly devoting attention to AI, elaborating on its potential uses, opportunities and threats.

    “This review indicates that more academics as well as practitioners, are increasingly devoting their attention to AI as they elaborate about its potential uses, as well as on its opportunities and threats.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    IEEE introduced free access to AI Ethics and Governance standards on 17 January 2023.

    “On the 17th January 2023, IEEE introduced free access to AI Ethics and Governance standards.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites IEEE, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    AI dialogue formats enable AI to admit their mistakes.

    “Their dialogue formats enable them to respond to questions (Thorp, 2023), to admit their mistakes (Barrot, 2023), and to even reject requests (Crawford & Paglen, 2021), if they are not recognized as appropriate.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Barrot, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    AI dialogue formats enable AI to reject requests if they are not recognized as appropriate.

    “Their dialogue formats enable them to respond to questions (Thorp, 2023), to admit their mistakes (Barrot, 2023), and to even reject requests (Crawford & Paglen, 2021), if they are not recognized as appropriate.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Crawford & Paglen, 2021

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    IEEE standards advocate the importance of transparency (of autonomous systems and of employer data governance) as well as of data privacy.

    “The standards advocate the importance of transparency (of autonomous systems and of employer data governance) as well as of data privacy.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites IEEE, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    IEEE SA provides free access to its global socio-technical standards to guide practitioners to engage in trustworthy AI innovations.

    “Currently, IEEE Standards Association (IEEE SA) provides free access to its global socio-technical standards to guide practitioners to engage in trustworthy AI innovations.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites IEEE, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    AI innovations can interact with online users through two-way communications.

    “AI innovations can interact with online users through two-way communications (Camilleri & Troise, 2023).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Camilleri & Troise, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    AI advancements are raising serious concerns among businesses and their stakeholders including governments, academia and civil societies, regarding the risk of possible harm associated with the use of intelligent, learning technologies.

    “Very often researchers discuss about how the advancements of AI are raising serious concerns among the businesses themselves and their stakeholders including the governments, academia and civil societies, regarding the risk of possible harm associated with the use intelligent, learning technologies (Galaz et al., 2021;John-Mathews et al., 2022).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Galaz et al., 2021

      Not checked in the free run.

    • cites John-Mathews et al., 2022

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    IEEE standards make reference to the United Nations Convention on the Rights of the Child.

    “In fact, they make reference to the United Nations Convention on the Rights of the Child.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites IEEE, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    The AI Bill of Rights is a voluntary guideline intended to protect the interests of American citizens who will be using AI innovations.

    “Their 'AI Bill of Rights' is a voluntary guideline that is intended to protect the interests of American citizens who will be using AI innovations.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites WhiteHouse, 2022

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Many companies are increasingly relying on AI algorithms before making strategic decisions.

    “Many companies are increasingly relying on AI algorithms, prior to making strategic decisions (Janiesch et al., 2021;Rąb-Kettler & Lehnervp, 2019).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Janiesch et al., 2021

      Not checked in the free run.

    • cites Rąb-Kettler & Lehnervp, 2019

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Several companies use ML/DL algorithms for business process automation, fraud prevention, malware detection, spam filtering, and predictive maintenance of recommender systems.

    “Several companies are using ML/DL algorithms for business process automation (BPA), fraud prevention, malware detection, spam filtering, as well as for the predictive maintenance of recommender systems, among other purposes (Engel et al., 2022;Romao et al., 2019).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Engel et al., 2022

      Not checked in the free run.

    • cites Romao et al., 2019

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    The EU's Artificial Intelligence Act includes accuracy, clear and adequate information, detailed documentation, high quality datasets that reduce risks and discrimination, human oversight measures, logging of activities to trace any tampering of data, robustness, and security.

    “European Union (EU)'s Artificial Intelligence Act (EU, 2021) Accuracy; Clear and adequate information; Detailed documentation; High quality datasets that reduce risks and discrimination; Human oversight measures; Logging of activities to trace any tampering of data; Robustness; Security.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites EU, 2021

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    The AI Bill of Rights states that automated systems are expected to explain how and when AI is being used and should provide clear information on how it works.

    “Their automated systems are expected to explain how and when AI is being used and should provide clear information on how it works.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites WhiteHouse, 2022

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    AI automated decisions may be incorrect or not factual because their algorithms are not always up to date.

    “Its automated decisions may be incorrect, or not factual, as their algorithms are not always up to date.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Core ideas behind responsible AI, such as bias prevention, transparency and fairness, are already aligned with fundamental principles of CSR.

    “They implied that core ideas behind responsible AI, such as bias prevention, transparency and fairness, are already aligned with fundamental principles of CSR.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Renieris et al., 2022

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    The researcher relied on a rigorous systematic review of articles about AI governance.

    “It explains that the researcher relied on a rigorous systematic review of articles about AI governance.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Organizations' social licence to operate paradigm expects them to justify corporate decisions and behaviours, including about responsible AI governance, with stakeholders including policy makers.

    “This argumentation is also related to the organizations' social licence to operate paradigm (Camilleri, 2017), as they are expected to justify corporate decisions and behaviours including about responsible AI governance with stakeholders including policy makers, among others.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Camilleri, 2017

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationnone given

    There are still challenges for business practitioners to keep pace with the latest developments in the real world and with the proliferation of rich data that could or could not be uploaded into AI systems.

    “However, in reality, there are still a number of challenges for business practitioners to keep up the pace with the latest developments in the real world and with the proliferation of rich data that could/could not be uploaded into AI systems (for various reasons).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    One IEEE standard is focused on evaluating the effects of autonomous and intelligent technologies on all citizens, including children.

    “In addition, one of IEEE standards is focused on evaluating the effects of autonomous and intelligent technologies on all citizens, including children.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites IEEE, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    ISO 26000's non-binding principles encourage organizations to engage in ethical behaviours, respect the rule of law, respect international norms of behaviour, and respect human rights.

    “Its non-binding principles encourage them to engage in ethical behaviours, respect the rule of law, respect international norms of behaviour and to respect human rights (Camilleri, 2019).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Camilleri, 2019

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    What is learned by AI systems may quickly become obsolete or outdated and could no longer be applicable in the near future.

    “What is/was learned by AI systems may quickly become obsolete or outdated and could no longer be applicable in the near future (Thorp, 2023).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Thorp, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    As of May 2019, OECD started raising awareness about its principles that guide practitioners in the creation of innovative and trustworthy AI systems.

    “As of May 2019, OECD has started raising awareness about its principles that guide practitioners in the creation of innovative and trustworthy AI systems.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites OECD, 2019

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    AI dialogue formats enable AI to respond to questions.

    “Their dialogue formats enable them to respond to questions (Thorp, 2023), to admit their mistakes (Barrot, 2023), and to even reject requests (Crawford & Paglen, 2021), if they are not recognized as appropriate.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Thorp, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    ML/DL applications in business intelligence and analytics can identify important information in datasets and reveal patterns, trends, cycles and anomalies from big data and small data.

    “Other ML/DL applications are related to business intelligence (BI) and analytics, as algorithms can be used to identify important information in datasets, and reveal patterns, trends, cycles and anomalies from the big data as well as from small data (Carvalho et al., 2019).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Carvalho et al., 2019

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    The AI Bill of Rights clarifies that AI designers, developers, and deployers have to ensure they safeguard users' data privacy and protect them from algorithmic discrimination.

    “It clarifies that AI designers, developers, and deployers have to ensure that they are committed to safeguard their users' data privacy and to protect them from algorithmic discriminations.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites WhiteHouse, 2022

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    The AI Bill of Rights states that users ought to be able to opt out when they want and to communicate with a human customer service agent to resolve queries or find solutions.

    “Users ought to be in a position to opt out, when they want, and to communicate with a human customer service agent to resolve queries or to find solutions to their problems.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites WhiteHouse, 2022

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    The AI Bill of Rights identifies five principles meant to guide practitioners in the development and utilization of AI designs.

    “In October 2022, American policy makers released a document that identified five principles that are meant to guide practitioners in the development and utilization of AI designs.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites WhiteHouse, 2022

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    DL algorithms can process complex patterns in texts, images, audio and video, and can provide reliable insights and predictions into the future.

    “They can complex patterns in texts, images, audio and video, and can provide reliable insights and predictions into the future (Buhmann & Fieseler, 2023).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Buhmann & Fieseler, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Microsoft methodically quantifies risks to minimize the time to remediation of predictable or known failures and to avoid mistakes that may result in harm to human beings.

    “It methodically quantifies the risks to minimize the time to remediation of predictable or known failures and to avoid mistakes that may result in any harm to human beings.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Microsoft, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    AI governance is based on formal rules (legislative acts and binding regulations) and voluntary principles intended to guide practitioners in research, development and maintenance of AI systems.

    “AI governance is based on formal rules (including legislative acts and binding regulations) as well as on voluntary principles that are intended to guide practitioners in their research, development and maintenance of AI systems (Butcher & Beridze, 2019;Gonzalez et al., 2020).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Butcher & Beridze, 2019

      Not checked in the free run.

    • cites Gonzalez et al., 2020

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    OECD has dedicated a standard to transparent reporting and disclosures of AI processes.

    “OECD has also dedicated a standard to transparent reporting and disclosures of AI processes.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites OECD, 2019

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductreasoning

    The SLR procedures are clearly outlined so other scholars can replicate and validate the results reported in this paper.

    “Therefore, they could easily replicate and validate the results reported in this paper.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Camilleri et al., 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S1 (asserted), S3 (load-bearing). May combine two claims.

    Premisecitation

    Microsoft adds that its secure features increase the reliability of data and protect personal data from being disseminated with other users.

    “It adds that its secure features increase the reliability of data and protect personal data from being disseminated with other users.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Microsoft, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    IBM's foundational properties of its AI ethics rest on five pillars: explainability, fairness, robustness, transparency, and privacy.

    “It clarified that its foundational properties of its AI ethics rest on five pillars: (i) explainability (AI designs that deliver seamless experiences); (ii) fairness (AI designs that assist humans in making fairer choices); (iii) robustness (AI designs that are employed to make crucial decisions); (iv) transparency (AI designs that reinforce trust through disclosures); and, (v) privacy (AI designs”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites IBM, 2022

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    OECD specifies that all AI actors are expected to ensure that all of their procedures can be traced, to reduce the risks of AI systems.

    “OECD specifies that all AI actors are expected to ensure that all of their procedures can be traced, to reduce the risks of AI systems.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites OECD, 2019

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    IBM uses technology to implement guardrails at each stage of the AI/ML lifecycle, during data collection and in its instrumenting processes.

    “IBM uses technology to implement guardrails at each stage of the AI/ML lifecycle, during data collection and in its instrumenting processes.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites IBM, 2022

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Microsoft is accountable with its customers and partners about how its AI systems are impacting the world, in order to make informed choices.

    “The technology company is accountable with its customers and partners about how its AI systems are impacting the world, in order to make informed choices.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Microsoft, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    IBM recognized the importance of articulating governance policies based on principles, regulations and legislation, which support it in its AI strategic management and operations.

    “The multinational technology corporation recognized the importance of articulating governance policies based on principles, regulations and legislation, which are supporting it in its AI strategic management and operations.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites IBM, 2022

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    AI governance represents a regulatory framework that can support AI practitioners in strategy formulation and day-to-day operations.

    “Essentially, it represents a regulatory framework that can support AI practitioners in their strategy formulation and in day-to-day operations (Erdélyi & Goldsmith, 2022;Mullins et al., 2021;Schneider et al., 2022).”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Erdélyi & Goldsmith, 2022

      Not checked in the free run.

    • cites Mullins et al., 2021

      Not checked in the free run.

    • cites Schneider et al., 2022

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    The rationale behind responsible AI governance is to ensure that automated systems including ML/DL technologies support individuals and organizations in achieving long-term objectives while safeguarding all stakeholders' interests.

    “The rationale behind responsible AI governance is to ensure that automated systems including ML/DL technologies, are supporting individuals and organizations in achieving their long terms objectives, whist safeguarding the interests of all stakeholders (Corea et al., 2022;Hickok, 2022).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Corea et al., 2022 source

      Not checked in the free run.

    • cites Hickok, 2022

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    IBM is transparent with its reporting of AI activities for the benefit of different stakeholders.

    “It is also transparent with its reporting of AI activities for the benefit of different stakeholders.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites IBM, 2022

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    IBM dedicated a website to explain its guiding values and governance principles related to AI applications and processes.

    “IBM dedicated a website to explain its guiding values and governance principles related to AI applications and processes.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites IBM, 2022

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S1 (asserted), S3 (load-bearing).

    Premisecitation

    Microsoft suggests that its AI systems are inclusive as they empower everyone around the world, making sure that no one is left out.

    “It suggests that its AI systems are inclusive as they empower everyone around the world, making sure that no one is left out.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Microsoft, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Microsoft clarifies that members of minority communities are involved in the research, development and testing of AI designs and solutions.

    “It clarifies that members of minority communities are involved in the research, development and testing of AI designs and solutions.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Microsoft, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Microsoft's Framework for Responsible AI underlines that its systems are intended to treat diverse people in a fair manner by reducing existing stereotypes, cultural denigration, under-representation and bias.

    “Microsoft's Framework for Responsible AI underlines that its systems are intended to treat diverse people in a fair manner, by reducing existing stereotypes, cultural denigration, under-representation and bias.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Microsoft, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    IBM's AI governance framework is intended to operationalize AI with confidence through lifecycle governance, to manage risk and reputation, to strengthen regulatory compliance and to meet stakeholder demands.

    “Its AI governance framework is intended to operationalize AI with confidence through lifecycle governance, to manage risk and reputation, to strengthen regulatory compliance and to meet stakeholder demands.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites IBM, 2022

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Microsoft posits that it is transparent with stakeholders as it is clear about the strengths and limitations of its AI systems.

    “Microsoft posits that it is transparent with stakeholders as it is clear about the strengths and limitations of its AI systems.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Microsoft, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    OECD standards promote accountability, transparency and explainability; robustness, security and safety; fairness and human centered values; and sustainable development and inclusive growth.

    “Its standards promote accountability, transparency and explainability; robustness, security and safety; fairness and human centered values; as well as sustainable development and inclusive growth.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites OECD, 2019

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisereasoning

    OECD implies that everyone ought to be accountable for their actions.

    “It implies that everyone ought to be accountable for their actions.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites OECD, 2019

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Microsoft's AI systems are subject to ongoing monitoring and evaluation processes to manage and maintain extant systems, improve them over time, troubleshoot problematic issues and identify new uses.

    “Its AI systems are subject to ongoing monitoring, and evaluation processes to manage and maintain extant systems, to improve them over time, troubleshoot problematic issues and to identify new uses.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Microsoft, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Microsoft makes specific reference to AI security aspects including data origin and lineage, internal and external data usage, data corruption considerations, anomaly detections, and changes in data that might indicate users trying to acquire data.

    “It makes specific reference to AI security aspects including to data origin and lineage, internal and external data usage; data corruption considerations, anomaly detections, changes in the data that might indicate that there may be users who are trying to acquire data.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Microsoft, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Microsoft confirms that it is committed to protect the privacy of its users.

    “Moreover, the technology giant confirms that is committed to protect the privacy of their users.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Microsoft, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    OECD's AI principles urge practitioners to respect human rights and democratic values in all stages of their research and development.

    “OECD's AI principles urge practitioners to respect human rights and democratic values, in all stages of their research and development.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites OECD, 2019

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    The systematic literature review methodology used in this study ensures that the findings are rigorous and trustworthy.

    “This rigorous methodology ensures that the findings of this research are rigorous and trustworthy, as other scholars can follow the procedures that are clearly outlined in this SLR (Camilleri et al., 2023).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Camilleri et al., 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Microsoft's corporate website suggests that its AI governance is based on responsible dimensions, including accountability, transparency, reliability and safety, privacy and security, fairness and inclusiveness.

    “The company's corporate website suggests that its AI governance is based on responsible dimensions, including accountability, transparency, reliability and safety, privacy and security, fairness and inclusiveness.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Microsoft, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Microsoft assures its customers that it regularly evaluates operational factors of its AI systems to determine whether they are performing reliably and safely.

    “Microsoft assures its customers that it regularly evaluates operational factors of its AI systems, to determine whether they are performing reliably and safely.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Microsoft, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Microsoft reported that its AI products are designed to provide a similar quality of service for various demographic segments in society including marginalized groups in order to minimize disparities among different people.

    “It reported that its AI products are designed to provide a similar quality of service for various demographic segments in society including to marginalized groups in order to minimize disparities among different people.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Microsoft, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Premisenone given

    According to several data privacy laws, organizations are bound to inform users that they are gathering and storing their biometric data, and businesses are not authorized to use consumers' data without consent.

    “According to several data privacy laws that have been enacted in different jurisdictions, organizations are bound to inform users that they are gathering and storing their biometric data.The businesses that employ such technologies are not authorized to use their consumers' data without their consent.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

    Interpretationnone given

    Innovative clusters and network developments may result in better AI systems and can decrease the chances of possible risks.

    “Innovative clusters and network developments may result in better AI systems and can also decrease the chances of possible risks.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Previous research reported that AI's bias may result in inequality, discrimination and loss of jobs.

    “Previous research reported that AI's bias may result in inequality, discrimination and in the loss of jobs (Butcher & Beridze, 2019).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Butcher & Beridze, 2019

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Organizational leaders are expected to follow ethical norms, values and standards.

    “Moreover, they are expected to follow ethical norms, values and standards (Koniakou, 2023).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Koniakou, 2023

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Explainable AI systems might improve consumer trust and may enable engineers to develop other AI models by tracking provenance, ensuring reproducibility, and enabling checks and balances.

    “Explainable AI systems might improve consumer trust and may enable engineers to develop other AI models, as they are in a position to track provenance of every process, to ensure reproducibility, and to enable checks and balances (Schneider et al., 2022).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Schneider et al., 2022

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    AI models can encounter incidental adversities leading to data corruption or intentional adversities from hackers, and in both cases the AI model will be compromised and can result in system malfunctions.

    “There may be instances when AI models could encounter incidental adversities, leading to the corruption of data.Alternatively, they might encounter intentional adversities when they experience sabotage from hackers.In both cases, the AI model will be compromised and can result in system malfunctions (Papagiannidis et al., 2023).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Papagiannidis et al., 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Premisecitation

    Practitioners can be in a better position if they cooperate with stakeholders for the development of trustworthy AI and if they increase their human capacity to improve the quality of their intellectual properties.

    “Indeed, practitioners can be in better position if they cooperate with stakeholders for the development of trustworthy AI and if they increase their human capacity to improve the quality of their intellectual properties (Camilleri et al., 2023).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Camilleri et al., 2023

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    These companies are voluntarily self-regulating themselves as they promote accountability, fairness, privacy and robust AI systems.

    “Evidently, these companies, among others, are voluntarily self-regulating themselves as they promote accountability, fairness, privacy and robust AI systems.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Interpretationreasoning

    Implementing AI governance procedures allows businesses to minimize unnecessary risks and safeguard the well-being of all stakeholders.

    “Ultimately, it is in their interest to operate their company in a responsible manner, and to implement AI governance procedures.This way they can minimize unnecessary risks and safeguard the well-being of all stakeholders.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Interpretationreasoning

    It is in companies' interest to continuously engage with internal stakeholders (and employees) and to educate them about AI governance dimensions intended to promote accountable, transparent, explainable, interpretable, reproducible, fair, inclusive and secure AI solutions.

    “Arguably, it is in their interest to continuously engage with internal stakeholders (and employees), and to educate them about AI governance dimensions, that are intended to promote accountable, transparent, explainable interpretable reproducible, fair, inclusive and secure AI solutions.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    Explainability of AI systems could improve their transparency, trustworthiness and accountability, and reduce bias and unfairness.

    “Arguably, the explainabilty of AI systems could improve their transparency, trustworthiness and accountability.At the same time, it can reduce bias and unfairness.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    Many researchers identified risks of AI, warning that AI could disseminate misinformation, foster prejudice, bias and discrimination, raise privacy concerns, and lead to loss of jobs.

    “Many researchers identified some of the risks of AI (Li et al., 2021;Magas & Kiritsis, 2022).In many cases, they warned that AI could disseminate misinformation, foster prejudice, bias and discrimination, raise privacy concerns, and could lead to the loss of jobs (Butcher & Beridze, 2019).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Li et al., 2021

      Not checked in the free run.

    • cites Magas & Kiritsis, 2022

      Not checked in the free run.

    • cites Butcher & Beridze, 2019

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Interpretationreasoning

    By engaging and educating stakeholders, companies could maximize AI benefits and minimize their risks as well as associated costs.

    “Hence, they could maximize AI benefits, minimize their risks as well as associated costs.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Companies are expected to communicate about their data privacy policies with their target audiences.

    “Companies are expected to communicate about their data privacy policies with their target audiences (Wong, 2020).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Wong, 2020

      Not checked in the free run.

80 of 80

  • The paper relies on this without support.

    There is still limited research focused on AI principles and regulatory guidelines for the developers of expert systems like machine learning (ML) and/or deep learning (DL) technologies.

    “However, for the time being, there is still limited research focused on AI principles and regulatory guidelines for the developers of expert systems like machine learning (ML) and/or deep learning (DL) technologies.”

  • The paper relies on this without support.

    There are limited contributions focused on AI governance frameworks that provide substantive (outcome-based) and reflexive (process-based) guidelines to practitioners who are developing AI innovations.

    “For the time being, there are limited contributions that are focused on AI governance frameworks that provide substantive (outcome-based) and reflexive (process-based) guidelines to practitioners who are developing AI innovations.”

  • The paper relies on this without support.

    More individuals and organizations are utilizing AI systems for different applications.

    “This research raises awareness on the importance of AI governance in an age where more individuals and organizations are utilizing AI systems for different applications.”

  • The paper relies on this without support.

    Various governments and international organizations are stepping in with their commitment to protect their citizens and the businesses' interests.

    “Various governments and international organizations are stepping in with their commitment to protect their citizens and the businesses' interests.”

  • The paper relies on this without support.

    Several regulatory authorities are outlining governance principles and guidelines that are intended to support practitioners in the development of AI, ML and DL technologies, with the aim to mitigate and reduce the risks associated with them.

    “As a result, several regulatory authorities are outlining governance principles and guidelines that are intended to support practitioners in the development of AI, ML and DL technologies, with the aim to mitigate and reduce the risks associated with them.”

  • The paper relies on this without support.

    These disruptive AI technologies may be used in an irresponsible manner and/or for malicious purposes.

    “Arguably, these disruptive AI technologies may be used in an irresponsible manner and/or for malicious purposes.”

  • The paper relies on this without support.

    The adoption of these disruptive AI technologies could raise serious concerns of different stakeholders in society.

    “Hence, their adoption could raise serious concerns of different stakeholders in society.”

  • The paper relies on this without support.

    The European Union put forward a proposed AI regulatory framework called the Artificial Intelligence Act in April 2021.

    “The European Union (EU) put forward its proposed AI regulatory framework that is referred to 'The Artificial Intelligence Act (AI Act)' in April 2021.”

  • The paper relies on this without support.

    The AI Act specifies objectives to increase the safety and security of AI systems, facilitate investment in automated systems, reinforce responsible AI governance through regulations and principles, and create a trustworthy and safe ecosystem for AI development.

    “It specifies that its objectives are: (i) to increase the safety and security of AI systems, as they have to respect relevant legislation on fundamental rights and should reflect EU values; (ii) facilitate investment in automated systems; (iii) reinforce responsible AI governance through regulations and principles; (iv) to create a trustworthy and safe eco-system for the development of AI systems.”

  • The paper relies on this without support.

    The EU Commission developed a risk-based approach pyramid that identifies four levels of risk: minimal, limited, high-risk, and unacceptable risk.

    “The EU Commission developed a risk-based approach pyramid that identifies four levels of risk: (i) minimal risk, (ii) limited risk, (iii) high-risk, and (iv) unacceptable risk.”

  • The paper relies on this without support.

    On 25 May 2022, the Singaporean government launched AI.Verify, an AI Governance Testing Framework and Toolkit for companies that may want to prove that their AI systems are responsible and trustworthy.

    “On the 25 May 2022, the Singaporean government launched AI.Verify, an AI Governance Testing Framework and Toolkit for companies that may want to prove that their AI systems are responsible and trustworthy.”

  • The paper relies on this without support.

    Google, Meta and Microsoft, among other businesses, have already adopted the Singaporean framework to confirm their AI governance credentials.

    “Google, Meta and Microsoft among other businesses have already adopted the Singaporean framework, to confirm their AI governance credentials.”

  • The paper relies on this without support.

    The Singaporean model framework integrated additional considerations like robustness and reproducibility to increase its relevance and usability.

    “Subsequently, the model framework integrated additional considerations like robustness (to conform with IBM's AI governance principles) and reproducibility, in order to increase its relevance and usability.”

  • The paper relies on this without support.

    It is difficult to understand how AI systems, including deep learning models and their neural networks, learn during training phases.

    “In reality, it is difficult to understand how AI systems, including deep learning models and their neural networks are learning (as they acquire, process and store data) during training phases.”

  • The paper relies on this without support.

    AI systems are often considered black box models.

    “They are often considered as black box models.”

  • The paper relies on this without support.

    AI systems learn from the data fed to them, and skewed or implicitly biased data may result in inappropriate outputs.

    “In many cases, AI systems are learning from the data that is fed to them.If the data are skewed and/or if they comprise implicit bias into them, they may result in inappropriate outputs.”

  • The paper relies on this without support.

    Consumers are increasingly concerned about the privacy of their data.

    “Consumers are increasingly concerned about the privacy of their data.”

  • The paper relies on this without support.

    AI-enabled products are increasingly capturing and storing large quantities of consumer information.

    “AI-enabled products, including dialogue systems like chatbots and virtual assistants, as well as digital assistants (e.g., like Siri, Alexa or Cortana), and/or wearable technologies such as smart watches and sensorial smart socks, among others, are increasingly capturing and storing large quantities of consumer information.”

  • The paper relies on this without support.

    In many cases, customers are not aware that they are sharing personal information with AI-enabled products.

    “In many cases, the customers are not aware that they are sharing personal information to them.”

  • The paper relies on this without support.

    AI algorithms are vulnerable to cyberattacks by malicious actors.

    “AI algorithms are vulnerable to cyberattacks by malicious actors.”

  • The paper relies on this without support.

    Routine testing can increase data protection, improve security levels and minimize the risks of incidents.

    “Routine testing can increase data protection, improve security levels and minimize the risks of incidents.”

  • The paper relies on this without support.

    Policy makers often face a 'pacing problem' because innovation proliferates faster than legislation, and governments tend to be reactive in implementing regulatory interventions relating to innovations.

    “Arguably, one has to bear in mind that, in many cases, policy makers have to face a 'pacing problem' as the proliferation of innovation is much quicker than legislation.As a result, governments tend to be reactive in the implementation of regulatory interventions relating to innovations.”

  • The paper relies on this without support.

    Currently, there is no universal definition on AI fairness.

    “Currently, there is no universal definition on AI fairness.”

  • The paper relies on this without support.

    Many companies including Facebook, Google, IBM and Microsoft are striving to involve diverse staff and become inclusive and culturally aware to minimize bias in AI processes.

    “For this reason, many companies including Facebook, Google, IBM and Microsoft, among others are striving in their endeavours to involve members of staff hailing from diverse backgrounds.These technology conglomerates are trying to become as inclusive and as culturally aware as possible in order to minimize bias from affecting their AI processes.”

  • The paper relies on this without support.

    Many legislators are striving to pressurize AI actors to describe algorithms, yet publication of algorithms is useless if outsiders cannot access the data of the AI model.

    “Many legislators are striving in their endeavours to pressurize AI actors to describe the algorithms they use in automated decision-making, yet the publication of algorithms is useless if outsiders cannot access the data of the AI model.”

  • The paper relies on this without support.

    Governments regulate business and industry to establish technical, safety and quality standards and to monitor compliance with them.

    “The governments' policy is to regulate business and industry to establish technical, safety and quality standards as well as to monitor their compliance.”
26 of 26

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?

  • When AI systems influence or partially automate sustainability-related decisions, the allocation of responsibility between managers, organizations, and algorithmic systems becomes increasingly opaque.

    “Recent research emphasizes that when AI systems influence or partially automate sustainability-related decisions, such as ESG risk assessment, supplier selection, or investment prioritization, the allocation of responsibility between managers, organizations, and algorithmic systems becomes increasingly opaque [26,27].”

    From Artificial Intelligence in Managerial Decision-Making for Sustainable Business Models: A Systematic Literature Review (2026) · cites this paper as [27]

    The Background section says companies rely on AI for strategic decisions, and §4.1 (Accountability and transparency; page unavailable) describes AI models as difficult to understand and calls for developer accountability. The paper does not say that responsibility becomes increasingly opaque between managers, organizations, and algorithms when AI influences sustainability decisions, nor does it discuss the cited ESG, supplier-selection, or investment-prioritization examples in that context. (LLM, conf=0.82)

    Experimental — accuracy not yet measured

Cites

Flagged first, then by citations

Research works printed but not linked: 10

  1. Bertino E., 2021, AI for security and security for AI. In proceedings of the eleventh ACM conference on data and application security and privacy, 333
  2. EU, 2021, Regulation of the European Parliament and the council laying down harmonized rules on artificial intelligence (artificial intelligence act) and amending certain union legislative acts
  3. IBM, 2022, Introducing IBM AI governance: IBM AI governance is a new, one‐stop solution built on IBM cloud Pak® for data
  4. IEEE, 2023, IEEE introduces new program for free access to AI ethics and governance standards
  5. Microsoft, 2023, Responsible and trusted AI
  6. Nation S., 2019, National Artificial Intelligence Strategy: Advancing our smart nation journey
  7. OECD, 2019, Recommendation of the council on artificial intelligence
  8. Renieris E. M., 2022, Should organizations link responsible AI and corporate social responsibility?It's Complicated
  9. 10.48550/arXiv.2011.10672 · DOI
  10. WhiteHouse, 2022, Blueprint for an AI bill of rights: Making automated systems work for the American people

Cited by

Sorted by citations

Related

Authors
Camilleri, M. A.
Affiliations
Northwestern University (US)
Corresponding author
Camilleri, Mark Anthony
ORCID
Camilleri, Mark Anthony
Venue
Expert Systems
Publisher
Wiley
Published
2023-07-18
Volume
41
Issue
7
ISSN
  • 0266-4720 (ISSN-L)
  • 1468-0394
DOI
10.1111/exsy.13406
OpenAlex
W4384827671
Cited by
331
References
85
Citations by year
2026: 173 · 2025: 112 · 2024: 44 · 2023: 2
FWCI
82.89
Type
journal article
Language
en
Open access
Licence
Locations
  • Expert Systems (publishedVersion, cc-by)
  • OAR@UM (University of Malta) (submittedVersion, other-oa)
Topics
  • Social Sciences › Safety Research › Ethics and Social Impacts of AI
  • Social Sciences › Law › Digitalization, Law, and Regulation
  • Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
Keywords
AI governance, corporate social responsibility, AI ethics, transparency and accountability, reproducibility, privacy, AI safety, cybersecurity risks, AI regulation, machine learning, deep learning, stakeholder responsibility

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.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-08.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-08.

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