Hostname: page-component-cd9895bd7-gxg78 Total loading time: 0 Render date: 2024-12-23T09:31:49.708Z Has data issue: false hasContentIssue false
  • English
  • Français

Liability Rules for AI-Related Harm: Law and Economics Lessons for a European Approach

Published online by Cambridge University Press:  16 September 2022

Shu Li Open the ORCID record for Shu Li [Opens in a new window] ,
Michael Faure Open the ORCID record for Michael Faure [Opens in a new window]  and
Katri Havu
Shu Li*
Affiliation:
University of Helsinki, Faculty of Law, Porthania, Yliopistonkatu 3, Helsinki, 00100, Finland
Michael Faure
Affiliation:
Maastricht University, METRO, Bouillonstraat 1–3, Maastricht, 6211 LH, The Netherlands Erasmus University Rotterdam, Rotterdam, Zuid-Holland, The Netherlands
Katri Havu
Affiliation:
University of Helsinki, Faculty of Law, Porthania, Yliopistonkatu 3, Helsinki, 00100, Finland
*
*Corresponding author. Email: [email protected]
Get access Rights & Permissions [Opens in a new window]

Abstract

The potential of artificial intelligence (AI) has grown exponentially in recent years, which not only generates value but also creates risks. AI systems are characterised by their complexity, opacity and autonomy in operation. Now and in the foreseeable future, AI systems will be operating in a manner that is not fully autonomous. This signifies that providing appropriate incentives to the human parties involved is still of great importance in reducing AI-related harm. Therefore, liability rules should be adapted in such a way to provide the relevant parties with incentives to efficiently reduce the social costs of potential accidents. Relying on a law and economics approach, we address the theoretical question of what kind of liability rules should be applied to different parties along the value chain related to AI. In addition, we critically analyse the ongoing policy debates in the European Union, discussing the risk that European policymakers will fail to determine efficient liability rules with regard to different stakeholders.

Keywords

AI-related harmartificial intelligencedeterrencedeveloperslaw and economicsliability rulesoperatorsrisk-bearing
Type
Articles
Information
European Journal of Risk Regulation , Volume 13 , Issue 4 , December 2022 , pp. 618 - 634
Copyright
© The Author(s), 2022. Published by Cambridge University Press

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 Council Directive 85/374/EEC of 25 July 1985 on the approximation of the laws, regulations and administrative provisions of the Member States concerning liability for defective products [1985] OJ L 210.

2 See P Machnikowski, “Producers’ Liability in the EC Expert Group Report on Liability for AI” (2020) 11(2) Journal of European Tort Law 137, 139.

3 See, eg, European Commission High-Level Expert Group on Artificial Intelligence, A Definition of AI: Main Capacities and Disciplines (2019) p 6; P Čerka, J Grigienė and G Sirbikytė, “Liability for Damages Caused by Artificial Intelligence” (2015) 31(3) Computer Law & Security Review 376; SJ Russell and P Norvig, Artificial Intelligence: A Modern Approach (4th edition, London, Pearson 2021) p 19.

4 Commission, “White Paper on Artificial Intelligence” COM(2020) 65 final, 10–11.

5 R Koulu, “Human Control over Automation: EU Policy and AI Ethics” (2020) 12 European Journal of Legal Studies 9.

6 See, eg, IN Cofone, “Servers and Waiters: What Matters in the Law of AI” (2018) 21 Stanford Technology Law Review 167; JS Gordon, “Artificial Moral and Legal Personhood” (2020) 36 AI & Society 457; SM Solaiman, “Legal Personality of Robots, Corporations, Idols and Chimpanzees: A Quest for Legitimacy” (2017) 25(2) Artificial Intelligence and Law 155. Some commentators have also proposed, from a legal standpoint, to treat AI systems as agents of their controllers: R Calo, “Robots as Legal Metaphors” (2016) 30 Harvard Journal of Law & Technology 209.

7 See, eg, European Expert Group on Liability and New Technologies – New Technologies Formation (EG-NTF), Report: Liability for Artificial Intelligence and Other Emerging Digital Technologies (2019) p 37.

8 See Commission, supra, note 4, 15.

9 M Buiten, A de Streel and M Peitz, “EU Liability Rules for the Age of Artificial Intelligence” (2021) <https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3817520> (last accessed 10 January 2022).

10 F Pasquale, The Black Box Society (Cambridge, MA, Harvard University Press 2015) pp 8–9.

11 J Burrell, “How the Machine ‘Thinks’: Understanding Opacity in Machine Learning Algorithms” 3(1) Big Data & Society 1.

12 MF Grady, “Unavoidable Accident” (2009) 5(1) Review of Law & Economics 177.

13 Commission, supra, note 4, 6.

14 E Talley, “Automatorts: How Should Accident Law Adapt to Autonomous Vehicles? Lessons from Law and Economics” (2019) Hoover Institution Working Group on Intellectual Property, Innovation, and Prosperity <https://www.hoover.org/sites/default/files/ip2-19002-paper.pdf> (last accessed 10 January 2022).

15 RH Coase, “The Problem of Social Cost” (1960) 3 The Journal of Law & Economics 1.

16 WY Oi, “The Economics of Product Safety: A Rejoinder” (1974) 5(2) The Bell Journal of Economics and Management Science 689.

17 GL Priest, The Rise of Law and Economics: An Intellectual History (Abingdon-on-Thames, Routledge 2020) p 43.

18 VP Goldberg, “Toward an Expanded Economic Theory of Contract” 10(1) Journal of Economic Issues 45.

19 S Shavell, Foundations of Economic Analysis of Law (Cambridge, MA, Harvard University Press 2004) p 175.

20 European Parliament, “The resolution of 16 February 2017 with recommendations to the Commission on Civil Law Rules on Robotics” (2015/2103(INL)).

21 Commission, “Artificial Intelligence for Europe” SWD(2018) 137 final.

22 Commission, supra, note 4, 12.

23 ibid, 21.

24 Commission, Report on the safety and liability implications of Artificial Intelligence, the Internet of Things and robotics, COM(2020).

25 ibid, 12 (emphasis added).

26 EG-NTF, supra, note 7, 5, Key finding [3].

27 ibid, 42, Key finding [13]. This viewpoint was reiterated by members of EG-NTF in a report drafted by the European Law Institute (ELI). See BA Koch et al, “Response of the European Law Institute: Public Consultation on Civil Liability Adapting Liability Rules to the Digital Age and Artificial Intelligence” (2022) 13(1) Journal of European Tort Law 25.

28 European Parliament, “Resolution of 20 October 2020 with recommendations to the Commission on a civil liability regime for artificial intelligence (2020/2014(INL))” (2020) <https://www.europarl.europa.eu/doceo/document/A-9-2020-0178_EN.html> (last accessed 10 January 2022).

29 ibid, para 8.

30 ibid, Arts 4 and 8.

31 Commission, “Proposal for a Regulation of the European Parliament and of the Council Laying down Harmonised Rules on Artificial Intelligence (Artificial Intelligence Act) and Amending Certain Union Legislative Acts”, COM(2021) 206 final.

32 See Commission, Inception Impact Assessment, Ares (2021)4266516 <https://ec.europa.eu/info/law/better-regulation/have-your-say/initiatives/12979-Civil-liability-adapting-liability-rules-to-the-digital-age-and-artificial-intelligence_en> (last accessed 25 June 2022).

33 Commission, “Adapting Civil Liability Rules to the Digital Age and Artificial Intelligence: Factual summary report on public consultation” (2022), Ref. Ares(2022)2620305 <https://ec.europa.eu/info/law/better-regulation/have-your-say/initiatives/12979-Civil-liability-adapting-liability-rules-to-the-digital-age-and-artificial-intelligence/public-consultation_en> (last accessed 25 June 2022).

34 See Machnikowski, supra, note 2; C Wendehorst, “Strict Liability for AI and Other Emerging Technologies” (2020) 11(2) Journal of European Tort Law 150; BA Koch, “Liability for Emerging Digital Technologies: An Overview” (2020) 11(2) Journal of European Tort Law 115.

35 Koch et al, supra, note 27.

36 See, eg, B Schütte, L Majewski and K Havu, “Damages Liability for Harm Caused by Artificial Intelligence – EU Law in Flux” (2021) Helsinki Legal Studies Research Paper 69 <https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3897839> (last accessed 10 January 2022).

37 A Bertolini and F Episcopo, “The Expert Group’s Report on Liability for Artificial Intelligence and Other Emerging Digital Technologies: A Critical Assessment” (2021) 12(3) European Journal of Risk Regulation 1.

38 S Shavell, Economic Analysis of Accident Law (Cambridge, MA, Harvard University Press 2009) p 3.

39 ibid.

40 Shavell, supra, note 19, 267–69.

41 Shavell, supra, note 38, 9–10.

42 S Shavell, “Strict Liability versus Negligence” (1980) 9(1) The Journal of Legal Studies 1.

43 JP Brown, “Toward an Economic Theory of Liability” (1973) 2(2) The Journal of Legal Studies 323.

44 Shavell, supra, note 38, 26.

45 G Dari-Mattiacci, “On the Definitions of Care and Activity Level and the Choice of Liability Rules” (2003) <https://dspace.library.uu.nl/bitstream/1874/723/24/c3.pdf> (last accessed 10 January 2022).

46 Shavell, supra, note 38, 30–31.

47 The literature also discusses the possibility of sharing the residual risk between tortfeasors and victims. However, that cannot result in an optimal outcome either. See E Carbonara, A Guerra and F Parisi, “Sharing Residual Liability: The Cheapest Cost Avoider Revisited” (2016) 45(1) The Journal of Legal Studies 173.

48 G Dari-Mattiacci and F Parisi, “Liability Rules: An Economic Taxonomy” in M Bussani and AJ Sebok (eds), Comparative Tort Law (Cheltenham, Edward Elgar Publishing 2021) pp 112–32.

49 S Shavell, “The Mistaken Restriction of Strict Liability to Uncommon Activities” (2018) 10 Journal of Legal Analysis 1.

50 Shavell, supra, note 19, 257.

51 Shavell, supra, note 38, 186.

52 S Shavell, “On Liability and Insurance” (1982) 13(1) The Bell Journal of Economics 120.

53 MG Faure, “Alternative Compensation Mechanisms as Remedies for Uninsurability of Liability” (2004) 29(3) The Geneva Papers on Risk and Insurance: Issues and Practice 455.

54 Shavell, supra, note 19, 267.

55 MA Geistfeld, Principles of Product Liability (St Paul, MN, Foundation Press Thomson/West 2011) p 11.

56 As a result, whenever damage was caused by a defective product, consumers could only claim against the seller who had a direct contractual relationship with them. Later, the seller had to claim against the party further up the supply chain until the claim went to the manufacturer. This approach could be very costly and time-consuming. See M Ebers, A Janssen and O Meyer, “Comparative Report” in M Ebers, A Janssen and O Meyer (eds), European Perspectives on Producers’ Liability: Direct Producer’s liability for Non-conformity and the Seller’s Right of Redress (Cologne, Otto Schmidt/De Gruyter European Law Publishers 2009) p 4.

57 Some of the influential American cases reflecting this trend are MacPherson v. Buick Motor Co., 217 N.Y. 382, 111 N.E. 1050 (N.Y. 1916); Escola v. Coca Cola Bottling Co., 24 Cal.2d 453, 150 P.2d 436 (Cal. 1944); Henningsen v. Bloomfield Motors, Inc., 32 N.J. 358, 161 A.2d 69 (N.J. 1960) and Greenman v. Yuba Power Products, Inc., 59 Cal.2d 57, 27 Cal. Rptr. 697, 377 P.2d 897 (Cal. 1963).

58 L Kaplow, “Optimal Deterrence, Uninformed Individuals, and Acquiring Information About Whether Acts Are Subject to Sanctions” (1990) 6 Journal of Law, Economics, & Organization 93.

59 JCP Goldberg and BC Zipursky, “The Easy Case for Products Liability Law: A Response to Professors Polinsky and Shavell” (2010) 123(8) Harvard Law Review 1919.

60 KN Hylton, “The Law and Economics of Products Liability” 88 Notre Dame Law Review 2457.

61 GL Priest, “The Invention of Enterprise Liability: A Critical History of the Intellectual Foundations of Modern Tort Law” (1985) 14(3) The Journal of Legal Studies 461.

62 MG Faure, “Economic Analysis of Product Liability” in P Machnikowski (ed.), European Product Liability: An Analysis of the State of the Art in the Era of New Technologies (Cambridge, Intersentia 2016) p 619.

63 See, eg, S Lohsse, R Schulze and D Staudenmayer, “Liability for Artificial Intelligence” in S Lohsse, R Schulze and D Staudenmayer (eds), Liability for Artificial Intelligence and the Internet of Things (Baden-Baden, Nomos Verlagsgesellschaft mbH & Co. KG 2019) pp 16–17.

64 G Howells, “Protecting Consumer Protection Values in the Fourth Industrial Revolution” (2020) 43(1) Journal of Consumer Policy 145.

65 See also, eg, Koch et al, supra, note 27, 34.

66 See also G Howells, C Twigg-Flesner and C Willett, “Product Liability and Digital Products” in TE Synodinou et al (eds), EU Internet Law (Berlin, Springer 2017).

67 See, eg, Koch, supra, note 34, 121–22.

68 Commission, supra, note 24, 6–7.

69 M Wansley, “The End of Accidents” (2021) 55 U.C. Davis Law Review 269.

70 W Kowert, “The Foreseeability of Human–Artificial Intelligence Interactions” (2017) 96 Texas Law Review 181.

71 This does not necessarily mean that the developers of AI systems should be placed under the product liability regime. Whether they are regarded as producers or they will be subject to a different liability regime based on strict liability is a matter of policy choice. The question of institutional design will be touched upon in Section V.

72 See, eg, MU Scherer, “Regulating Artificial Intelligence Systems: Risks, Challenges, Competencies, and Strategies” (2015) 19 Harvard Journal of Law & Technology 353.

73 MG Faure, L Visscher and F Weber, “Liability for Unknown Risks – A Law and Economics Perspective” (2016) 7(2) Journal of European Tort Law 198.

74 See also, eg, A Galasso and H Luo, “Tort Reform and Innovation” (2017) 60(3) The Journal of Law and Economics 385.

75 G Spindler, “User Liability and Strict Liability in the Internet of Things and for Robots” in S Lohsse, R Schulze and D Staudenmayer (eds), Liability for Artificial Intelligence and the Internet of Things (Baden-Baden, Nomos Verlagsgesellschaft mbH & Co. KG 2019) pp 125, 140–41.

76 See also, eg, BA Koch, “The ‘European Group on Tort Law’ and Its ‘Principles of European Tort Law’” (2005) 53(1) The American Journal of Comparative Law 189, 200–01.

77 See also, eg, G Wagner, “Robot Liability” in S Lohsse, R Schulze and D Staudenmayer (eds), Liability for Artificial Intelligence and the Internet of Things (Baden-Baden, Nomos Verlagsgesellschaft mbH & Co. KG 2019) p 50.

78 G Wagner, “Liability for Artificial Intelligence: A Proposal of the European Parliament” <https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3886294> pp 17–18 (last accessed 11 December 2021).

79 See, eg, Bertolini and Episcopo, supra, note 37.

80 See, eg, NA Smuha, “Beyond the Individual: Governing AI’s Societal Harm” (2021) 10(3) Internet Policy Review 1.

81 MG Faure and S Li, “Artificial Intelligence and (Compulsory) Insurance” (2022) 13(1) Journal of European Tort Law 1.

82 See also Bertolini and Episcopo, supra, note 37.

83 Wendehorst, supra, note 34.

84 O Dheu, C Ducuing and P Valcke, “The Emperor’s New Clothes: A Roadmap for Conceptualizing the New Vehicle” (2020) 75 TRANSIDIT 12.

85 See also, eg, Spindler, supra, note 75, 127.

86 See also Koch et al, supra, note 27, 39.

87 ibid, 26.

88 See also BA Koch, “Product Liability 2.0 – Mere Update or New Version?” in S Lohsse, R Schulze and D Staudenmayer (eds), Liability for Artificial Intelligence and the Internet of Things (Baden-Baden, Nomos Verlagsgesellschaft mbH & Co. KG 2019); see also MF Lohmann, “Liability Issues Concerning Self-Driving Vehicles” (2016) 7(2) European Journal of Risk Regulation 335.

89 See also JA Pattinson, HB Chen and S Basu, “Legal Issues in Automated Vehicles: Critically Considering the Potential Role of Consent and Interactive Digital Interfaces” (2020) 7(1) Humanities and Social Sciences Communications 1.

90 See, eg, Buiten et al, supra, note 9.

91 Machnikowski, supra, note 2, 139.

92 Art. 2 of the PLD.

93 For the attitudes of Member States regarding the issue of whether digital goods are qualified as “products”, see Commission, Evaluation of Council Directive 85/374/EEC on the approximation of laws, regulations and administrative provisions of the Member States concerning liability for defective products: Final Report (2018) 69–72 <https://op.europa.eu/en/publication-detail/-/publication/d4e3e1f5-526c-11e8-be1d-01aa75ed71a1/language-en> (last accessed 10 January 2022).

94 See Schütte et al, supra, note 36, 23. There is also literature that suggests that the strict liability of developers should not rely on the notion of “defect”. See Spindler, supra, note 75, 136.

95 Art 7(e) of the PLD. Note that the Member States had the possibility to exclude the defence (Art 15(1)(b)) and therefore the availability of the defence currently varies across the EU.

96 EG-NTF, supra, note 7, 6, Key finding [14].

97 ibid, 42.

98 European Parliament, supra, note 28, Art 4(1).

99 ibid, para 8.

100 See also, eg, Spindler, supra, note 75, 137.

101 Wagner, supra, note 77, 34–36.

102 P Fabian et al, “Reasonable, Adequate and Efficient Allocation of Liability Costs for Automated Vehicles: A Case Study of the German Liability and Insurance Framework” (2018) 9(3) European Journal of Risk Regulation 548.

103 See also, eg, regarding the discussion on rules versus standards, I Ehrlich and RA Posner, “An Economic Analysis of Legal Rule-Making” (1974) 3 Journal of Legal Studies 257; L Kaplow, “Rules versus Standards: An Economic Analysis” (1992) 42(3) Duke Law Journal 557; F Weber, “European Integration Assessed in the Light of the Rules vs. Standards Debate” (2013) 35(2) European Journal of Law and Economics 187.

104 Commission, “Proposal for a Regulation of the European Parliament and of the Council Laying down Harmonised Rules on Artificial Intelligence (ARTIFICIAL INTELLIGENCE ACT) and Amending Certain Union Legislative Acts”, COM(2021) 206 final, Annexes 2 & 3.

105 For example, while AVs could reduce some risks caused by the tiredness of drivers, they may generate risks that drivers could easily avoid. See Wagner, supra, note 77, 44.

106 ibid, 46.

107 Wendehorst, supra, note 34, 173.

108 Wagner, supra, note 77, 49.

109 Lohsse et al, supra, note 63, 18; see also Wagner, supra, note 78, 18–19.

110 EG-NTF, supra, note 7, 39, Key finding [11].

111 E Karner, “Liability for Robotics: Current Rules, Challenges, and the Need for Innovative Concepts” in S Lohsse, R Schulze and D Staudenmayer (eds), Liability for Artificial Intelligence and the Internet of Things (Baden-Baden, Nomos Verlagsgesellschaft mbH & Co. KG 2019) pp 117, 122; see also Bertolini and Episcopo, supra, note 79.

112 For example, according to SAE International, all on-road motor vehicles are categorised into six levels in accordance with their level of automation. Correspondingly, the degree of control exercised by frontend operators is also different. The liability of frontend operators must be adjusted to the degree of control. The extent to which a frontend operator is subject to which liability, however, can only be decided in a sector-by-sector manner. Soft laws, such as standards, can help crystallise the concrete role taken by a frontend operator. See SAE, J3016 Visual Chart <https://www.sae.org/blog/sae-j3016-update> (last accessed 10 January 2022).

113 See, eg, European Parliament, supra, note 28; Commission, supra, note 24.

114 See Commission, supra, note 24, 12.

115 See, eg, European Parliament, supra, note 28.

116 See also, eg, Koch et al, supra, note 27.

117 EG-NTF, supra, note 7, 39, Key finding [11].

118 European Parliament, supra, note 28, Art 4(4).

119 Faure and Li, supra, note 81.