CoChat
Sign Up
Cited by 156Digitalization, Law, and RegulationEthics and Social Impacts of AIArtificial Intelligence in Law

On the Governance of Artificial Intelligence through Ethics Guidelines

Open access
LARSSON, Stefan·2020·Asian Journal of Law and Society
DOI 10.1017/als.2020.19
Open this paper in CoChat — free accountOpen PDF

Abstract

This article uses a socio-legal perspective to analyze the use of ethics guidelines as a governance tool in the development and use of artificial intelligence (AI). This has become a central policy area in several large jurisdictions, including China and Japan, as well as the EU, focused on here. Particular emphasis in this article is placed on the Ethics Guidelines for Trustworthy AI published by the EU Commission’s High-Level Expert Group on Artificial Intelligence in April 2019, as well as the White Paper on AI, published by the EU Commission in February 2020. The guidelines are reflected against partially overlapping and already-existing legislation as well as the ephemeral concept construct surrounding AI as such. The article concludes by pointing to (1) the challenges of a temporal discrepancy between technological and legal change, (2) the need for moving from principle to process in the governance of AI, and (3) the multidisciplinary needs in the study of contemporary applications of data-dependent AI.

AI snapshot

Written from the paper’s full text.

TL;DR

This article uses a socio-legal perspective to analyze ethics guidelines as a governance tool for artificial intelligence, focusing on the EU’s Ethics Guidelines for Trustworthy AI (April 2019) and White Paper on AI (February 2020). It concludes that AI governance faces a temporal gap between technological and legal change, needs to move from principle to process, and requires multidisciplinary research.

Key findings

  • The article argues that AI’s definitional instability—no clear consensus across public, computer-science, and legal contexts—shapes governance, and that contemporary AI mainly includes machine learning, natural language processing, image recognition, neural networks, and deep learning.
  • The EU High-Level Expert Group, consisting of 52 members, defines trustworthy AI as legal, ethical, and robust; it sets out four ethical principles (respect for human autonomy, prevention of harm, fairness, explicability) and seven prerequisites, including human agency and oversight, technical robustness and safety, privacy and data governance, transparency, diversity/nondiscrimination/fairness, societal and environmental wellbeing, and accountability.
  • A 2019 global study identified 84 documents containing ethical principles or guidelines for AI and found relative unanimity on five principles: transparency, justice and fairness, non-harmful use, responsibility, and integrity/data protection, with transparency the single most common; over 350 organizations tested the EU assessment list during 2019.
  • The EU White Paper proposes a risk-based, sector-specific approach distinguishing high-risk applications, initially including healthcare, transport, energy, and parts of the public sector; the article also briefly compares China’s AIDP (2017), Made in China 2025 (2015), and Beijing AI Principles (May 2019) with Japan’s Social Principles of Human-Centric AI (March 2019).

Methods

  • Socio-legal and conceptual analysis of ethics guidelines as a governance tool, with particular focus on EU documents: the Ethics Guidelines for Trustworthy AI (April 2019), the White Paper on AI (February 2020), and related High-Level Expert Group definition and policy documents.
  • Draws on secondary literature and policy documents, including a 2019 global landscape study of 84 AI ethics documents and comparisons to Chinese and Japanese AI governance initiatives.
  • No empirical data collection or sample; the analysis is interpretive and argumentative, concluding with three socio-legal statements on temporality, process, and multidisciplinary research needs.

Limitations

  • The paper states that the Chinese and Japanese developments “will only briefly be addressed here,” so the Asian comparison is explicitly limited.
  • Evident from the text: it is a conceptual/socio-legal analysis without primary empirical data, systematic coding, or measured outcomes; its conclusions are normative and interpretive.
  • Evident from the text: the analysis focuses mainly on EU documents and selected secondary sources, so its findings may not generalize to other jurisdictions or implementation contexts.

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

Integrity signals

Registry

Crossref · 10.1017/als.2020.19

Cites

51 of 51 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 31 cited sources checked · 0 supported · 0 flagged

  • Premisecitation

    Ethics Guidelines as a governance tool have seen very strong development over the last few years.

    “Ethics Guidelines in particular, as a governance tool, have seen very strong development over the last few years.”
    Quote in the paperNot a substantial claim
  • Premisecitation

    The study concluded there is relative unanimity globally on at least five principle-based approaches: transparency, justice and fairness, non-harmful use, responsibility, and integrity/data protection.

    “The study concluded that there is relative unanimity globally on at least five principle-based approaches of ethical character: (1) transparency, (2) justice and fairness, (3) non-harmful use, (4) responsibility, and (5) integrity/data protection.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [28]

      Not checked in the free run.

  • Interpretationreasoning

    The way in which the standpoints in principle are intended to be implemented is often unclear.

    “In particular, the way in which the standpoints in principle are intended to be implemented is often unclear.”
    Quote in the paperSubstantial claim
  • Interpretationcitation

    The governance using ethics guidelines is highly dependent on recent insights from critical AI studies about societal challenges relating to fairness, accountability, and transparency.

    “The governance using ethics guidelines is highly dependent on recent insights from critical AI studies about the societal challenges relating to fairness, accountability, and transparency.16”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [16] source

      Not checked in the free run.

  • Interpretationreasoning

    Given aspects of learning in data-dependent AI, there is an interaction in which human values and societal structures constitute the training data.

    “Given aspects of learning in data-dependent AI, there is an interaction at hand in which human values and societal structures constitute the training data.”
    Quote in the paperSubstantial claim
  • Classified as probably not a substantial claim. 55
  • May combine two claims. 23

  • Classified as probably not a substantial claim. May combine two claims.

    Findingcitation

    In April 2019, the expert group published the Ethics Guidelines for Trustworthy Artificial Intelligence, which indicate issues of responsibility, transparency, and data protection as central parts of trustworthy AI.

    “In April 2019, the expert group published the Ethics Guidelines for Trustworthy Artificial Intelligence (hereinafter the Ethics Guidelines), 11 which—despite explicitly pointing out that the guidelines do not deal with legal issues—clearly indicate issues of responsibility, transparency, and data protection as entirely central parts of the development of trustworthy AI.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [11] source

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Findingreasoning

    The guidelines focus on ethical issues and robustness, but leave legal issues outside the explicit guidelines.

    “The guidelines focus on ethical issues (b) and robustness (c), but leave legal issues (a) outside the explicit guidelines.”
    Quote in the paperSubstantial claim

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

    Findingcitation

    For the realization of trustworthy AI, the expert group addresses seven main prerequisites that must be evaluated and managed continuously during the entire life-cycle of the AI system: human agency and oversight; technical robustness and safety; privacy and data governance; transparency; diversity, nondiscrimination, and fairness; societal and environmental wellbeing; and accountability.

    “However, for the realization of trustworthy AI, they address seven main prerequisites, which, they argue, must be evaluated and managed continuously during the entire life-cycle of the AI system: 1. Human agency and oversight 2. Technical robustness and safety 3. Privacy and data governance 4. Transparency 5. Diversity, nondiscrimination, and fairness 6. Societal and environmental wellbeing 7. A”
    Quote in the paperNot a substantial claim

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

    Findingcitation

    Although the guidelines emphasize ethics and robustness, not legality, anti-discrimination and protection of privacy are developed as two of the seven central ethical prerequisites for trustworthy AI.

    “As mentioned, although the guidelines emphasize that they focus on ethics and robustness, and not on issues of legality, it is interesting to note that both anti-discrimination (5) and protection of privacy (3) are developed as two of the seven central ethical prerequisites for the implementation of trustworthy AI.”
    Quote in the paperNot a substantial claim

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

    Findingreasoning

    The expert group highlights that new legal measures and control mechanisms may be needed to safeguard adequate protection against negative effects and to enable correct supervision and implementation.

    “They highlight that new legal measures and control mechanisms may be needed to safeguard adequate protection against negative effects, and to enable correct supervision and implementation.”
    Quote in the paperNot a substantial claim

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

    Findingcitation

    The expert group recommends a risk-based approach that is proportional and effective in guaranteeing that AI is legal, ethical, and robust in its adaptation to fundamental rights.

    “In relation to the investment and policy recommendations also published by the expert group, it recommends features such as a risk-based approach that is both proportional and effective in guaranteeing that AI is legal, ethical, and robust in its adaptation to fundamental rights.39”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [39] source

      Not checked in the free run.

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

    Findingnone given

    The expert group calls for comprehensive mapping of relevant EU regulations to assess the extent to which various regulations are still fulfilling their purposes in an AI-driven world.

    “Interestingly, the expert group calls for comprehensive mapping of relevant EU regulations to be carried out, in order to assess the extent to which the various regulations are still fulfilling their purposes in an AI-driven world.”
    Quote in the paperNot a substantial claim

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

    Interpretationcitation

    The High-Level Expert Group's definition of AI concentrates particularly on autonomy and includes both physical robots and pure software systems.

    “This definition concentrates particularly on autonomy—namely that there is a measure of agency in AI systems—and points out that the systems can consist of both physical robots and also pure software systems.”
    Quote in the paperNot a substantial claim

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

    Interpretationreasoning

    The ethics guidelines focus on normative standpoints but are often weak from a procedural perspective.

    “In other words, the ethics guidelines focus on normative standpoints, but are often weak from a procedural perspective.”
    Quote in the paperNot a substantial claim

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

    Interpretationcitation

    The definition of AI is in itself a form of conceptual control that impacts on the regulation debate, so we need to be careful and take a multidisciplinary approach when making definitions.

    “The definition is in itself a form of conceptual control that impacts on the regulation debate, and we therefore need to both be careful and take a multidisciplinary approach when making definitions.27”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [27] source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    The study establishes there are considerable differences in how these principles are interpreted, why they are considered important, what issue/domain/actors they relate to, and how they should be implemented.

    “At the same time, the study establishes that there are considerable differences in how these principles are interpreted; why they are considered important; what issue, domain or actors they relate to; and how they should be implemented.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [28]

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation

    Over the last few years, a number of ethics guidelines have been developed relating to AI by companies, research associations, and government representatives.

    “Over the last few years, a number of ethics guidelines have been developed relating to AI: by companies, research associations, and government representatives.12”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [12] source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    The examples in the High-Level Expert Group's definition indicate the governance objects of the Ethics Guidelines.

    “At the same time, the examples provide a clear indication of what they are aiming at and, extrapolating, what the governance objects of the Ethics Guidelines consist of.”
    Quote in the paperSubstantial claim

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

    Interpretationnone given

    Many ethics guidelines overlap in part with already-existing legislation, but it is often unclear how the legislation and guidelines are intended to interact more precisely.

    “Many overlap in part with already-existing legislation, but it is often unclear how the legislation and guidelines are intended to interact more precisely.”
    Quote in the paperNot a substantial claim

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

    Interpretationreasoning

    The High-Level Expert Group's definition, by emphasizing autonomy, can be interpreted as not applying to all drones or all linked-up devices-only to those with an autonomous or learning element.

    “As autonomy is emphasized, this can in combination be interpreted as not applying to all drones or all linked-up devices—only to those that have an autonomous or even learning element.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    Issues fairly well anchored in law, such as responsibility, anti-discrimination, and data protection, still fall within the framework for ethics.

    “It does this despite the fact that issues that are fairly well anchored in law, such as responsibility, anti-discrimination, and—not least—data protection, still fall within the framework for ethics.”
    Quote in the paperSubstantial claim

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

    Interpretationnone given

    The actual definition of AI is highly debatable and may depend on the disciplinary field, and it will arguably have an effect on its governance.

    “Since the actual definition of AI is highly debatable, and may depend on the disciplinary field in which the person making the definition is based, it will arguably have an effect on its governance.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    The Ethics Guidelines of the EU Commission's expert group are a clear sign of an ongoing governance challenge for the EU and its Member States.

    “The Ethics Guidelines of the EU Commission's expert group are a clear sign of an ongoing governance challenge for the EU and its Member States.”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    The intelligence concept is an explicit part of AI and is particularly elusive.

    “The High-Level Expert Group also notes that an explicit part of AI is the intelligence concept, which is a particularly elusive element that has been included since the area was originated.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    There are multidisciplinary advantages of not leaning too heavily towards a computer sciences-based definition of AI.

    “Finally, as a step towards a wider social sciences-based discussion and in light of the challenges that AI has displayed in its implementation and interaction with society's values and structure, it can be argued that there are multidisciplinary advantages of not leaning too heavily towards a computer sciences-based definition of AI.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    Even though ethical and robust AI is to some extent often already reflected in existing laws, its full implementation may reach beyond existing legal obligations.

    “Even though ethical and robust AI is to some extent often already reflected in existing laws, its full implementation may reach beyond existing legal obligations.”
    Quote in the paperSubstantial claim

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

    Premisecitation

    Legg and Hutter gathered more than 70 different definitions of the intelligence concept.

    “Legg and Hutter, for example, gather together more than 70 different definitions of the intelligence concept in itself.23”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [23]

      Not checked in the free run.

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

    Premisecitation

    A 2019 study identified 84 documents containing ethical principles or guidelines for AI.

    “For example, a study of the global AI ethics landscape, published in 2019, identified 84 documents containing ethical principles or guidelines for AI.28”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [28]

      Not checked in the free run.

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

    Premisecitation

    There is no clear consensus on how AI should best be defined.

    “Despite—or perhaps because of—the increased attention that AI and its developed methods are receiving in multidisciplinary research, media, and policy work, there is no clear consensus on how AI should best be defined.”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites [17] source

      Not checked in the free run.

    • cites [18] source

      Not checked in the free run.

    • cites [19] source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationnone given

    The governance issue must inevitably deal with temporal aspects of the difference between how legislation is formed and how rapid development has been for the underlying elements of AI.

    “At the same time, the governance issue must inevitably deal with temporal aspects of the difference between how legislation is formed and how rapid development has been for the underlying elements of AI.”
    Quote in the paperNot a substantial claim

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

    Interpretationcitation

    The White Paper's risk proposal is an either/or approach on risk, and more nuanced alternatives have been proposed elsewhere, for example by the German Data Ethics Commission.

    “This proposal is an either/or approach on risk, and more nuanced alternatives have been proposed elsewhere, for example by the German Data Ethics Commission.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [47]

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation

    The lack of transparency makes it difficult to identify and prove possible breaches of laws, including legal provisions that protect fundamental rights, attribute liability and meet the conditions to claim compensation.

    “The lack of transparency, the Commission brings forward, makes it "difficult to identify and prove possible breaches of laws, including legal provisions that protect fundamental rights, attribute liability and meet the conditions to claim compensation."”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [50] source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationnone given

    Member States' supervisory authorities and agencies will very likely be the ones to carry out relevant parts of any regulatory approach on AI focusing on what the High-Level Expert Group has expressed as a need for "explicability", that is auditability and explainability.

    “The Member States' supervisory authorities and agencies could be addressed specifically here too, in the sense that they will very likely be the ones to carry out relevant parts of any regulatory approach on AI focusing what the High-Level Expert Group has expressed as a need for "explicability"-that is, auditability and explainability.”
    Quote in the paperNot a substantial claim

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

    Interpretationreasoning

    The subsequent procedural stage is necessary both to strengthen the chances of implementing the principal values and to formalize them in legislation, assessment methodologies and standardization.

    “The subsequent procedural stage is necessary, however, both to strengthen the chances of implementing the principal values as well as to formalize in legislation, assessment methodologies, and standardization.”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    A recent study concludes that common elements of the Japanese and European notions of governance include that AI should be applied in a human-centric manner and should be committed to the fundamental (constitutional) rights of individuals and democracy.

    “a recent study concludes that common elements of both notions of governance include that AI should be applied in a manner that is "human-centric" and should be”
    • Also stated in: § 53. National Governance Committee for the New Generation Artificial Intelligence (2019) Governance principles of the New Generation Artificial Intelligence-developing responsible AI. · plain
      “committed to the fundamental (constitutional) rights of individuals and democracy.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [55]

      Not checked in the free run.

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

    Premisecitation

    According to Kozuka, Japan's Social Principles are more policy-oriented, while the European Ethics Guidelines have a rights-based approach.

    “A particular difference is, however, according to Kozuka, that Japan's Social Principles are more policy-oriented, while the European Ethics Guidelines have a rights-based approach.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [55]

      Not checked in the free run.

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

    Premisereasoning

    A key reason for the transparency the Ethics Guidelines call for is building and maintaining users' trust.

    “A key reason is to be building and maintaining users' trust.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationreasoning

    The lack of expressing power structures of contemporary data-driven platform markets, emphasized by critics, is perhaps a sign of the regulatory struggles to come in the leap from principles to process.

    “Perhaps the lack of expressing power structures of contemporary data-driven platform markets-emphasized by critics-is a sign of the regulatory struggles to come in the leap from principles to process.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Interpretationreasoning

    The great range of ethics guidelines displays a core of principal values but, being ethical guidelines, is relatively poor in procedural arrangements compared to law.

    “The great range of ethics guidelines still displays a core of principal values, but-being ethical guidelines-are relatively poor in procedural arrangements compared to law.”
    Quote in the paperSubstantial claim

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

    Premisenone given

    In May 2019 the Beijing AI Principles were released by the Beijing Academy of Artificial Intelligence, depicting the core of its AI development as the realization of beneficial AI for humankind and nature.

    “In May 2019, the so-called Beijing AI Principles, which is another set, were released by the Beijing Academy of Artificial Intelligence, depicting the core of its AI development as the realization of beneficial AI for humankind and nature.”
    Quote in the paperNot a substantial claim

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

    Interpretationcitation

    The notion that code substitutes for law as a mechanism of implementation is particularly meaningful in relation to automated policy implementation on large-scale digital platforms, which shapes both human and institutional behaviour.

    “This notion is particularly meaningful in relation to automated policy implementation on large-scale digital platforms, shaping both human and institutional behaviour.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [57]

      Not checked in the free run.

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

    Premisenone given

    In Japan, an expert group at the Japanese Cabinet Office elaborated the Social Principles of Human-Centric AI, published in March 2019 after public comments were solicited.

    “In Japan, an expert group at the Japanese Cabinet Office has elaborated on the Social Principles of Human-Centric AI (Social Principles), which was published in March 2019 after public comments were solicited.”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    The Commission hopes that a clear European regulatory framework would build trust among consumers and businesses in AI and thereby speed up the uptake of the technology.

    “The Commission's hope is that a clear European regulatory framework would "build trust among consumers and businesses in AI, and therefore speed up the uptake of the technology."”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [49]

      Not checked in the free run.

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

    Premisenone given

    The key element of a future regulatory framework for AI in Europe is to create a "unique ecosystem of trust," which is described as a policy objective in itself.

    “The key elements of a future regulatory framework for AI in Europe is to create a "unique ecosystem of trust," which is described as a policy objective in itself.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Interpretationnone given

    The difficulty of defining AI is one of the regulatory challenges that follow from the implementation and development of AI.

    “This article has put forward the difficulty of defining AI as one of the regulatory challenges that follow from the implementation and development of AI.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationcitation

    The speed of technological change is repeatedly a difficult challenge in relation to the inertia of traditional regulation, even though the "legal lag" is more complex than it may seem.

    “Even though the "legal lag" is more complex than it may seem, 58 the speed of change, in particular, is still repeatedly a difficult challenge in relation to the inertia of traditional regulation.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [59]

      Not checked in the free run.

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

    Premisenone given

    The White Paper makes a clear address to the human-centric approach based on the Communication on Building Trust in Human-Centric AI, which is also a central part of the Ethics Guidelines.

    “The White Paper makes a clear address to the human-centric approach based on the Communication on Building Trust in Human-Centric AI, which is also a central part of the Ethics Guidelines discussed above.”
    Quote in the paperNot a substantial claim

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

    Interpretationnone given

    There is a clear value-base in the White Paper, with a particular focus on the concept of trust.

    “There is a clear value-base in the White Paper, with a particular focus on the concept of trust”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    In 2019 a National Governance Committee for the New Generation Artificial Intelligence was established, which published a set of governance principles.

    “In 2019, a National Governance Committee for the New Generation Artificial Intelligence was established, which published a set of governance principles.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [53]

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Interpretationreasoning

    If the growth in ethics guidelines is regarded as an expression of the rapidity of the development of AI methods, the procedural stage is an expected second stage.

    “If one can regard the growth in ethics guidelines as an expression of the rapidity of the development of AI methods, the procedural stage is an expected second stage.”
    Quote in the paperNot a substantial claim

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

    Interpretationnone given

    The interpretation of existing legislation in light of the functionalities, possibilities and challenges of AI systems is a matter of serious concern associated with major challenges.

    “The interpretation of existing legislation in light of functionalities, possibilities, and challenges of AI systems is also a matter of serious concern associated with major challenges.”
    Quote in the paperNot a substantial claim

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

    Interpretationreasoning

    If the ethics guidelines are regarded as industry's reluctance to accept regulation, as a soft version of legislation intended to be toothless, then the procedural stage will meet considerable resistance.

    “However, if one regards the ethics guidelines as industry's reluctance to accept regulation of its activities, as a soft version of legislation that is intended to be toothless, then the procedural stage will meet considerable resistance.”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    A key proposal in the White Paper on AI is a risk-based, sector-specific approach to regulating AI in which high-risk applications are distinguished from all other applications.

    “A key proposal in the White Paper is taking a risk-based, sector-specific approach to regulating AI, in which high-risk applications are distinguished from all other applications.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [45]

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    The AI-development issue has come to take a value-based and ethics-focused development within the European administration, with a focus on trustworthiness and human-centric design.

    “In conclusion, the AI-development issue has come to take a value-based and ethicsfocused development within the European administration with a focus on trustworthiness and human-centric design.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Interpretationcitation

    Legislative processes aimed at learning technologies with increasing agency require reflection, critical studies and more knowledge in order to find the desirable societal balances between various interests.

    “Legislative processes aimed at learning technologies with increasing agency 60 require reflection, critical studies, and more knowledge in order to be able to find the desirable societal balances between various interests.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [60]

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    The conceptual fuzziness of AI creates a challenge for regulation and governance, even though the heterogeneous AI field provides favourable conditions for research and development.

    “While the historically visionary and contemporary heterogeneous AI field arguably provides favourable conditions for research and development, the conceptual fuzziness creates a challenge for regulation and governance.”
    Quote in the paperSubstantial claim

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

    Premisecitation

    Law can be an extremely dynamic and adaptive organism, as legal scholars such as Karl Renner, who analysed property laws of Western Europe's industrialization, teach.

    “At the same time, legal scholars such as Karl Renner, who analyzed property laws of Western Europe's industrialization, also teach that law can be an extremely dynamic and adaptive organism.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Renner

      Not checked in the free run.

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

    Premisecitation

    Kozuka concludes that the role of the law as a mechanism of implementation will shrink and be substituted by code as the use of AI becomes widespread.

    “Interestingly, Kozuka-with references to Lawrence Lessig in the paper-concludes that "the role of the law as a mechanism of implementation will shrink and be substituted by the code as the use of AI becomes widespread."”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [56]

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Interpretationreasoning

    It is conceivable that central parts of the Ethical Guidelines may be formalized with support for European and national legislation and regulation, focusing on the importance of an ecosystem of trust.

    “It is conceivable that central parts of the Ethical Guidelines may be formalized with support for European and national legislation and regulation, focusing on the importance of ("an ecosystem of") trust.”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    A goal expressed in the AIDP is to establish initial ethical norms, policies and regulations related to AI development in China by 2020, to be further codified by 2025.

    “For example, a goal expressed in the AIDP is to establish initial ethical norms, policies, and regulations related to AI development in China by 2020, to be further codified by 2025.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [52]

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation

    The Commission concludes that transparency, traceability and human oversight are not specifically covered under current legislation in many economic sectors.

    “Interestingly enough, the Commission concludes that those regarding transparency, traceability, and human oversight are not specifically covered under current legislation in many economic sectors.”
    • Also stated in: § 4.1 The Temporality Issue of Technology and Law · plain
      “not clearly covered-or understood-under current legislation in many economic sectors.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [50] source

      Not checked in the free run.

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

    Premisecitation

    The Ethics Guidelines require processes to be transparent in the sense that the capacities and purpose of AI systems should be openly communicated and decisions should, to the extent possible, be explainable to those directly and indirectly affected.

    “The Ethics Guidelines argue for the need for processes to be transparent in the sense that the capacities and purpose of AI systems should be "openly communicated, and decisions-to the extent possible-explainable to those directly and indirectly affected."”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [40]

      Not checked in the free run.

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

    Premisecitation

    The Beijing AI Principles have been supported by various elite Chinese universities and companies including Baidu, Alibaba and Tencent.

    “These Principles have been supported by various elite Chinese universities and companies including Baidu, Alibaba, and Tencent.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [54]

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    The White Paper signifies an ongoing process of evaluation towards where the principled take on AI governance expressed by a multitude of ethics guidelines can find a balanced formalization in law.

    “The White Paper signifies an ongoing process of evaluation towards where the principled take on AI governance expressed by a multitude of ethics guidelines can find a balanced formalization in law.”
    Quote in the paperSubstantial claim
59 of 59

  • The paper relies on this without support.

    Machine learning depends on large amounts of data or examples as its foundation.

    “which is not least underlined by the dependence of machine learning on large amounts of data or examples as its foundation.”

  • The paper relies on this without support.

    Societal use and dependency on AI and machine learning are increasing.

    “As societal use and dependency on AI and machine learning are increasing, society increasingly needs to understand any negative consequences and risks, how interests and power are distributed, and what needs exist for both legal and other types of governance.”

  • The paper relies on this without support.

    The intelligence concept leads to human associations such as feelings and self-awareness that are not a living part of current applied AI.

    “The intelligence concept also leads to a number of human associations, such as the ability to have feelings and self-awareness that cannot be said to be a living part of the methods and technologies that are causing the explosion of applied AI today, and thus not a central object for governance through ethics guidelines.”

  • The paper relies on this without support.

    The Ethics Guidelines have had a clear impact on the subsequent White Paper on AI.

    “The Ethics Guidelines have had a clear impact on the subsequent White Paper on AI from the EU Commission (see below) but it still remains to be seen what sort of importance and impact all of these sources will have on European AI development.”

  • The paper relies on this without support.

    The composition of the High-Level Expert Group has not avoided criticism.

    “The composition has not avoided criticism, however.”

  • The paper relies on this without support.

    When the Ethics Guidelines were published, they were criticized by members of the expert group itself.

    “When the Ethics Guidelines were published, they were also criticized by members of the expert group itself.”

  • The paper relies on this without support.

    The Commission's intention that all Member States should have their own AI strategies by the middle of 2019 did not completely materialize.

    “For example, the Commission expressed an intention that all Member States should have their own strategies in place by the middle of 2019, which did not completely materialize.”

  • The paper relies on this without support.

    The Chinese and Japanese developments on AI policy and governance are significant from an Asian socio-legal perspective.

    “From an Asian socio-legal perspective, the Chinese and the Japanese developments on AI policy and governance are significant but will only briefly be addressed here.”
8 of 8

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?

  • Companies that use data-driven decision-making are, on average, 6% more profitable and 5% more productive than their competitors.

    “It is estimated that companies that use data-driving decision-making are, on average, 6% more profitable and 5% more productive than their competitors.84”

    From Artificial Intelligence and Sustainable Decisions (2023) · cites this paper as [84]

Cites

Flagged first, then by citations

Research works printed but not linked: 5

  1. 2019, A Definition of AI: Main Capabilities and Disciplines: Definition Developed for the Purpose of the AI HLEG’s Deliverables
  2. 2017, New Generation Artificial Intelligence Development Plan (AIDP)
  3. Legg, 2007, Advances in Artificial General Intelligence: Concepts, Architectures and Algorithms, 157, 17
  4. 2019, Policy and Investment Recommendations for Trustworthy Artificial Intelligence
  5. 2019, Ethics Guidelines for Trustworthy AI

Cited by

Sorted by citations

Related

Authors
LARSSON, S.
Affiliations
Lund University (SE)
Corresponding author
LARSSON, Stefan
ORCID
LARSSON, Stefan
Venue
Asian Journal of Law and Society
Publisher
Cambridge University Press (CUP)
Published
2020-10
Volume
7
Issue
3
Pages
437-451
ISSN
  • 2052-9015 (ISSN-L)
  • 2052-9023
DOI
10.1017/als.2020.19
OpenAlex
W3028544364
Semantic Scholar
48268a9d24db58c9a5331acaef8821a7e983a321
Cited by
156
References
51
Citations by year
2026: 27 · 2025: 34 · 2024: 36 · 2023: 29 · 2022: 14 · 2021: 12 · 2020: 4
FWCI
96.47
Type
journal article
Language
en
Open access
Licence
Locations
  • Asian Journal of Law and Society (publishedVersion, cc-by)
  • Lund University Publications (Lund University) (submittedVersion, other-oa)
Funders
  • Lunds Universitet (2019-00198)
  • Vetenskapsrådet (2019-00198)
Topics
  • Social Sciences › Law › Digitalization, Law, and Regulation
  • Social Sciences › Safety Research › Ethics and Social Impacts of AI
  • Social Sciences › Political Science and International Relations › Artificial Intelligence in Law
Keywords
AI governance, ethics guidelines, trustworthy AI, AI legislation, socio-legal analysis, China AI policy

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