SpringerLink
Log in

Liability of AI

  • Chapter
  • First Online:
Toward a Conceptual Network for the Private Law of Artificial Intelligence

Part of the book series: Law, Governance and Technology Series ((ISDP,volume 51))

  • 784 Accesses

Abstract

Asimov’s Laws have been often taken as a starting point for reflection on liability of AI, one example being the EP resolution of 16 February 2017 on Civil Law Rules on Robotics. However, the analysis made herein shows that this way of proceeding is not effective. Furthermore, the mostly recommended model for liability of the producer of AI, or another person responsible for AI, or the AI itself, is a strict liability regime. In addition to reflecting on the issue of legally-relevant damage caused by AI, the chapter examines the problem associated with establishing a causative link between action of AI and damage and the questions of negligence and standard of conduct for AI. It argues that there are some nontrivial situations involving AI when fault liability could be useful. For this reason, it contains a proposal of some new understanding of legal culpability, when this concept is applied to AI. The proposal is consistent with the concepts proposed in the previous chapters, such as the concept of legal subjectivity of AI, the capability for juridical actions of AI, the equivalent of free will and the discernment of AI.

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Log in via an institution

Subscribe and save

Springer+ Basic
EUR 32.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or Ebook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

Chapter
EUR 29.95
Price includes VAT (Germany)
  • Available as PDF
  • Read on any device
  • Instant download
  • Own it forever
eBook
EUR 128.39
Price includes VAT (Germany)
  • Available as EPUB and PDF
  • Read on any device
  • Instant download
  • Own it forever
Softcover Book
EUR 171.19
Price includes VAT (Germany)
  • Compact, lightweight edition
  • Dispatched in 3 to 5 business days
  • Free ship** worldwide - see info
Hardcover Book
EUR 171.19
Price includes VAT (Germany)
  • Durable hardcover edition
  • Dispatched in 3 to 5 business days
  • Free ship** worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Institutional subscriptions

Notes

  1. 1.

    There are different meanings of “responsibility” and different kinds of “liability”—see Hage (2017).

  2. 2.

    It is proposed in Resolution 2020 motive 8, that these problems should be faced. The European Parliament: “Considers that the Product Liability Directive (PLD) has, for over 30 years, proven to be an effective means of getting compensation […] but should nevertheless be revised to adapt it to the digital world […] urges the Commission to assess whether the PLD should be transformed into a regulation, to clarify the definition of ‘products’ by determining whether digital content and digital services fall under its scope and to consider adapting concepts such as ‘damage’, ‘defect’ and ‘producer’ […] the concept of ‘producer’ should incorporate manufacturers, developers, programmers, service providers as well as backend operators; calls on the Commission to consider reversing the rules governing the burden of proof for harm caused by emerging digital technologies […].”

  3. 3.

    Asimov (1976) and Anderson (2008).

  4. 4.

    Resolution 2017, point A.

  5. 5.

    Barbrook (2007).

  6. 6.

    Asimov (1985).

  7. 7.

    Asimov (1942), pp. 94–103.

  8. 8.

    See e.g. McCauley (2007).

  9. 9.

    The criteria of context may be useless in the case, when the situations very much alike externally are qualified: e.g. removal of the kidney being an effect of conscious consent and upon the legally valid rules vs. removal of the kidney without a conscious consent and illegally.

  10. 10.

    Remember the trolley dilemma first introduced by Foot (1967).

  11. 11.

    Murphy and Woods (2009).

  12. 12.

    Searle (1984) and Nilsson (2009).

  13. 13.

    EC High-Level Expert Group on AI (2019), A Definition of AI: Main Capabilities and Scientific Disciplines. Brussels.

  14. 14.

    Bostrom (2014).

  15. 15.

    Bostrom (2014). The concept of singularity was introduced by Vinge (2003) in the 1980s and then developed.

  16. 16.

    Murphy and Woods (2009).

  17. 17.

    Clarke (1994).

  18. 18.

    Williams (1973).

  19. 19.

    Lucas (1993).

  20. 20.

    See Open Letter to the European Commission. Artificial Intelligence and Robotics, 2019. http://www.robotics-openletter.eu, last access on the 4th of August 2022.

  21. 21.

    This way of thinking presents Proposal 2021.

  22. 22.

    The above ideas about the Asimov’s Law were presented in Księżak and Wojtczak (2020).

  23. 23.

    However, point (17) of Annex B includes an additional condition: “unless stricter national laws and consumer protection legislation is in force. The national laws of the Member States, including any relevant jurisprudence, with regard to the amount and extent of the compensation, as well as the limitation period, should continue to apply.”

  24. 24.

    A very clear juxtaposition of current national liability frameworks in EU is included in Evas (2020).

  25. 25.

    The problems of causation appearing when albeit perpetrator of harm is AI and some proposals of solving these problems are presented in Wojtczak and Księżak (2021).

  26. 26.

    Skinner (1969), p. 282; Nathan (2021), pp. 61–67.

  27. 27.

    Cf. Blanco-Justicia and Domingo-Ferrer (2019).

  28. 28.

    Cf. Wojtczak and Księżak (2021).

  29. 29.

    Before reporting a provider must establish a causal link. So, when someone reports, simultaneously admits that there is a causal link. Usually, the other party would have to proof the causal link at the court.

  30. 30.

    Stanicki et al. (2021).

  31. 31.

    O’Sullivan et al. (2019), p. 9.

  32. 32.

    Although there is certain difference: while it cannot be argued that medicine or equipment (e.g. respirator or dialyzer) could be substituted by a human doctor, it can be rationally said that a human doctor could substitute an AI or robot in cases where, historically, it was the AI or robots which substituted human doctors. So, the argument may sound like this: “The hospital is not liable for damage caused by not having the surgical robot, because it had the best surgeon among its staff and he did what he could.”

  33. 33.

    According to Article 2 point 1 of Regulation (EU) 2017/75 on medical devices which is applied from 26.05.2020 ‘medical device’ means any instrument, apparatus, appliance, software, implant, reagent, material or other article intended by the manufacturer to be used, alone or in combination, for human beings for one or more of the following specific medical purposes: “— diagnosis, prevention, monitoring, prediction, prognosis, treatment or alleviation of disease, — diagnosis, monitoring, treatment, alleviation of, or compensation for, an injury or disability, — investigation, replacement or modification of the anatomy or of a physiological or pathological process or state, — providing information by means of in vitro examination of specimens derived from the human body, including organ, blood and tissue donations, and which does not achieve its principal intended action by pharmacological, immunological or metabolic means, in or on the human body, but which may be assisted in its function by such means.”

    The following products shall also be deemed to be medical devices: “— devices for the control or support of conception; — products specifically intended for the cleaning, disinfection or sterilisation of devices as referred to in Article 1(4) and of those referred to in the first paragraph of this point.”

  34. 34.

    Regulation (EU) 2017/75 of the European Parliament and of the Council of 5 April 2017 on medical devices, amending Directive 2001/83/EC, Regulation (EC) No 178/2002 and Regulation (EC) No 1223 and repealing Council Directives 90/385/EEC and 93/42/EEC, L 117/1, https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX:32017R0745&from=PL, last access on the 4th of August 2022.

  35. 35.

    Article 22 section 1 GDPR: The data subject shall have the right not to be subject to a decision based solely on automated processing, including profiling, which produces legal effects concerning him or her or similarly significantly affects him or her.

  36. 36.

    The same will happen in many other areas, such as road traffic. At the beginning the autonomous vehicles will operate under additional control demanded by the law, but over time, the possibility of human driving will be restricted because it will obviously increase the danger of accident.

  37. 37.

    Hoeren and Niehoff (2018), pp. 308 and further.

  38. 38.

    Hoeren and Niehoff (2018), pp. 308 and further; Schoenberger (2019), pp. 171–203.

  39. 39.

    https://www.europarl.europa.eu/doceo/document/TA-8-2019-0081_EN.html, access on the 4th of August 2022.

  40. 40.

    https://airly.org/pl/alert-smogowy-kiedy-i-przy-jakich-wartosciach-oglaszany-jest-alarm-z-powodu-smogu/, last access on the 4th of August 2022.

  41. 41.

    The problem is similar to that relating to so-called conscience clauses, when the view of one person, e.g., a doctor or pharmacist, may prevail over the claim of another person to receive a certain service.

  42. 42.

    Of course, this problem is connected to political decisions and the concept of democracy accepted in a given time and a given place, but it does not mean that it should not be discussed or at least noticed.

  43. 43.

    https://www.jeder-mensch.eu/informationen/?lang=en, last access on the 4th of August 2022.

  44. 44.

    The rules proposed by EU bodies insist on the rule of risk-management and give many criteria which should be applied to assess the risk of harm posed by AI. The instance may be here Article 7 of Proposal 2021, which establishes the directives for the Commission regarding the method of updating the list in Annex III with the addition of high-risk AI systems, and Article 9 which details a risk management system.

  45. 45.

    Among others: Directorate-General for Internal Policies Policy Department for Citizen’s Rights and Constitutional Rights and Constitutional Affairs, Artificial Intelligence and Civil Liability. Study requested by the JURI Committee, PE 621.926—July 2020, Resolution 2017, Commission Staff Working Document: Liability for emerging digital technologies accompanying the document Communication from the Commission to the European Parliament, the European Council, the Council, the European Economic and Social Committee and the Committee of the Regions: Artificial Intelligence for Europe, Brussels, 25.4.2018, COM (2018) 237 final, Report from the Expert Group on Liability and New Technologies—New Technologies Formation, Liability for Artificial Intelligence and Other Emerging Digital Technologies, European Union 2019, Report from the Commission to the European Parliament, the Council and the European and Social Committee: Report on the safety and liability implications of Artificial Intelligence, the Internet of Things and Robotics, Brussels, 19.2.2020, COM (2020) 64 final, Resolution 2020, Hamon et al. (2020), EUR 30040.

  46. 46.

    Resolution 2020, Annex, B. (8): […] whoever creates, maintains, controls or interferes with the AI-system, should be accountable for the harm or damage that the activity, device or process causes. This follows from general and widely accepted liability concepts of justice, according to which the person who creates or maintains a risk for the public is liable if that risk causes harm or damage, and thus should ex-ante minimise or ex-post compensate that risk.

  47. 47.

    Gless et al. (2016) and Dafni (2018).

  48. 48.

    de-Wit et al. (2016).

  49. 49.

    Gazzaniga (2011), p. 222, cites R. Sapolsky, professor of neurology who said: “It is boggling that the legal system’s gold standard for an insanity defense – M’Naughten – is based on 166-year-old science. Our growing knowledge about the brain makes notions of volition, culpability, and, ultimately the very premise of the criminal justice system, deeply suspect”.

  50. 50.

    See the cases in Edersheim et al. (2012), also Greely (2011) and Weiss and Watson (2015).

  51. 51.

    Gazzaniga (2011), pp. 228, 252–253.

  52. 52.

    Mueller (1957–1958), Jordaan (2003) and de Maglie (2005).

  53. 53.

    There is huge literature on this issue, both technical and legal, see e.g. Blanco-Justicia and Domingo-Ferrer (2019), Brkan and Bonnet (2020) and Hamon et al. (2020).

  54. 54.

    Cf. Shapiro (2006) and DeGrazia (2006).

  55. 55.

    Neemeh (2018), p. 1.

  56. 56.

    Evans (2009).

  57. 57.

    Monterossi (2020), p. 5.

  58. 58.

    For the definition of radial category idealized cognitive models and composite prototype see Evans (2007), pp. 29, 104, 177–179.

  59. 59.

    Gobert (1994), p. 409.

  60. 60.

    The second of “General principles concerning the development of robotics and artificial intelligence for civil use” says that European Parliament: “2. Considers that a comprehensive Union system of registration of advanced robots should be introduced within the Union’s internal market where relevant and necessary for specific categories of robots, and calls on the Commission to establish criteria for the classification of robots that would need to be registered; in this context, calls on the Commission to investigate whether it would be desirable for the registration system and the register to be managed by a designated EU Agency for Robotics and Artificial Intelligence”.

  61. 61.

    Proving the improper or ineffective realization of a function is a matter of civil procedure. It may depend on the opinion of experts or the report of an entity controlling registration or certification.

  62. 62.

    It is analogical to Williams’ and Nagel’s moral luck. Williams (1982) and Nagel (1979).

References

Books and Articles

  • Anderson SL (2008) Asimov’s “Three Laws of Robotics” and machine metaethics. AI Soc 22(4):477–493. https://doi.org/10.1007/s00146-007-0094-5

    Article  Google Scholar 

  • Asimov I (1942) Roundaround. Astounding Science Fiction

    Google Scholar 

  • Asimov I (1976) Bicentennial man. Ballantine Books, New York

    Google Scholar 

  • Asimov I (1985) Robots and empire. Doubleday, New York

    Google Scholar 

  • Barbrook R (2007) Imaginary futures: from thinking machines to the global village. Pluto Press, London

    Google Scholar 

  • Beckers A, Teubner G (2021) The three liability regimes for artificial intelligence: algorithmic a cants, hybrids, crowds. Hart, Oxford

    Google Scholar 

  • Blanco-Justicia A, Domingo-Ferrer J (2019) Machine learning explain ability through comprehensible decision trees. In: Holzinger A, Kieseberg P, Min Tjoa A, Weippl E (eds) Machine learning and knowledge extraction. Springer, Cham. ISBN 978-3-030-29726-8

    Google Scholar 

  • Bostrom N (2014) Superintelligence: paths, dangers, Strategies. Oxford University Press, Oxford

    Google Scholar 

  • Brkan M, Bonnet G (2020) Legal and technical feasibility of the GDPR’s quest for explanation of algorithmic decisions: of black boxes, white boxes and fata morganas. Eur J Risk Regul. 11(18):II.2–II.3. ISSN 2190-8249

    Google Scholar 

  • Clarke R (1994) Asimov’s laws of robotics: implications for information technology. IEEE Comput 27(1):57–66

    Article  Google Scholar 

  • Dafni L (2018) Could AI agents be held criminally liable: artificial intelligence and the challenges for criminal law. South Carolina Law Rev 69:677. https://scholarcommons.sc.edu/cgi/viewcontent.cgi?article=4253&context=sclr, last access on the 4th of August 2022

  • De Maglie C (2005) Models of corporate criminal liability in comparative study. Washington Univ Global Stud Law Rev 4:547

    Google Scholar 

  • DeGrazia D (2006) On the question of Personhood beyond Homo sapiens. In: Singer P (ed) The defense of animals. Second Wave. Blackwell Publishing, Malden, pp 40–53

    Google Scholar 

  • de-Wit L, Alexander D, Ekroll V et al (2016) Is neuroimaging measuring information in the brain? Psychon Bull Rev 23:1415–1428. https://doi.org/10.3758/s13423-016-1002-0

    Article  Google Scholar 

  • Edersheim JG, Weintraub Brendel R, Price B (2012) Neuroimaging, diminished capacity and mitigation. In: Simpson JR (ed) Neuroimaging in forensic psychiatry. From clinic to the courtroom. Wiley and Sons

    Google Scholar 

  • Evans V (2007) A glossary of cognitive linguistics. Edinburgh University Press, Edinburgh

    Book  Google Scholar 

  • Evans EP (2009) The criminal prosecution and capital punishment of animals. The Lawbook Exchange Ltd., Clark, NJ

    Google Scholar 

  • Floridi, L, Cowls, J, Beltrametti, M, Chatila, R, Chazerand, P, Dignum, V, Luetge, Madelin R, Pagallo U, Rossi F, Schafer B, Valcke P, Vayena E (2018) “AI4People – An Ethical Framework for a Good AI Society: Opportunities, Risks, Principles and Recommendations”, forthcoming in Minds and Machines, December 2018. https://ssrn.com/abstract=3284141, last access on the 4th of August 2022

  • Foot P (1967) The problem of abortion and the Doctrine of the double effect. Oxford Rev 5

    Google Scholar 

  • Gazzaniga M (2011) Who is in charge? Free will and the science of the Brain. HarperCollins Publishers, Pymble (Australia)

    Google Scholar 

  • Gless S, Silverman E, Weigend T (2016) If Robot cause harm, who is to blame? Self-driving cars and criminal liability. New Crim Law Rev Int Interdiscip J 19(3):412–436

    Article  Google Scholar 

  • Gobert J (1994) Corporate criminality: four models of fault. Legal Stud 14(03). https://doi.org/10.1111/j.1748-121x.1994.tb0510.x

  • Greely HT (2011) Neuroscience and criminal responsibility: proving ‘Can’t Help Himself’ as a narrow bar to criminal liability. In: Freeman M (ed) Law and neuroscience. Current Legal Issues 2010 vol. 13. Oxford University Press, Oxford

    Google Scholar 

  • Hacker P, Krestel R, Grundmann S, Naumann F (2020) Explainable AI under contract and tort law: legal incentives and technical challenges. Artif Intell Law 228:415–439

    Article  Google Scholar 

  • Hage J (2017) Theoretical foundations for the responsibility of autonomous agents. Artif Intell Law 3(25)

    Google Scholar 

  • Hamon R, Junklewitz H, Sanchez I (2020) Research Centre Technical Report. Robustness and explainability of artificial intelligence – from technical to policy solutions. Publications Office of the European Union, Luxembourg. https://doi.org/10.2760/57493. (online), JRC119336

    Book  Google Scholar 

  • Hoeren T, Niehoff M (2018) Artificial intelligence in medical diagnoses and the right to explanation. Eur Data Prot Law Rev 4:3. https://doi.org/10.21552/edpl/2018/3/9

    Article  Google Scholar 

  • Jordaan L (2003) New perspective on the criminal liability of corporate bodies. Acta Juridica 48

    Google Scholar 

  • Kaplan J (2016) Artificial intelligence – what everyone needs to know. Oxford University Press, Oxford

    Google Scholar 

  • Księżak P, Wojtczak S (2020) Prawa Aismova, czyli science fiction jako fundament nowego prawa cywilnego. Forum Prawnicze 4(60). https://doi.org/10.32082/fp.v0i4(60).378

  • Lucas J (1993) Responsibility. A Clarendon Press Publication, Oxford

    Google Scholar 

  • McCauley L (2007) AI Armageddon and the Three Laws of Robotics. Ethics Inf Technol 9(2):153–164. https://doi.org/10.1007/s10676-007-9138-2

    Article  Google Scholar 

  • Monterossi MW (2020) Liability for the fact of autonomous artificial intelligence agents. Things, agencies and legal actors Global Jurist 20190054, eISSN 1934-2640, https://doi.org/10.1515/gj-2019-0054

  • Mueller GOW (1957–1958) Mens Rea and the corporation: a study of the Model Penal Code position on corporate criminal liability. Univ Pittsburgh Law Rev 19:21–50

    Google Scholar 

  • Murphy R, Woods D (2009) Beyond Asimov: the three laws of responsible robotics. IEEE Intell Syst 24(4):14–20. https://doi.org/10.1109/MIS.2009.69

    Article  Google Scholar 

  • Nagel T (1979) Moral luck. In: Nagel T (ed) Mortal questions. Cambridge University Press, Cambridge

    Google Scholar 

  • Nathan MJ (2021) Black Boxes: how science turns ignorance into knowledge. Oxford University Press

    Book  Google Scholar 

  • Neemeh ZA (2018) Husserlian empathy and simulationism, memphis: organization of phenomenological organizations VI: Phenomenology and Practical Life 2018, https://www.memphis.edu/philosophy/opo2019/pdfs/neemeh-zach.pdf, last access on the 4th of August 2022

  • Nilsson N (2009) The quest for artificial intelligence: a history of ideas and achievements. Cambridge University Press, New York

    Book  Google Scholar 

  • O’Sullivan S, Nevejans N, Allen C, Blyth A, Leonard S, Pagallo U, Holzinger K, Holzinger A, Sajid MI, Ashrafian H (2019) Legal, regulatory, and ethical frameworks for development of standards in artificial intelligence (AI) and autonomous robotic surgery. Int J Med Robot Comp Assisted Surg 15(1):1–12. https://doi.org/10.1002/rcs.1968

    Article  Google Scholar 

  • Schoenberger D (2019) Artificial intelligence in healthcare: a critical analysis of the legal and ethical implications. Int J Law Inf Technol 27:2

    Google Scholar 

  • Searle JR (1984) Minds, brains and science. Harvard University Press, Cambridge

    Google Scholar 

  • Shapiro P (2006) Moral agency in other animals. Theoret Med Bioeth 27(4):357–373

    Article  Google Scholar 

  • Skinner BF (1969) Contingencies of reinforcement: a theoretical analysis. Appleton-Century-Crofts, New York

    Google Scholar 

  • Stanicki P, Nowakowska K, Piwoński M, Żak K, Niedobylski S, Zaremba B, Oszczędłowski P (2021) The role of artificial intelligence in cancer diagnostics - a review. J Educ Health Sport 11(9):113–122. https://doi.org/10.12775/JEHS.2021.11.09.016

    Article  Google Scholar 

  • Vinge V (2003) Technological Singularity. http://cmm.cenart.gob.mx/delanda/textos/tech_sing.pdf, last access on the 4th of August 2022

  • Weiss KJ, Watson C (eds) (2015) Psychiatric expert testimony. Emerging applications. Oxford University Press, Oxford

    Google Scholar 

  • Williams B (1973) A critique of utilitarianism. In: Smart J, Williams B (eds) Utilitarianism for and against. Cambridge University Press, Cambridge

    Google Scholar 

  • Williams B (1982) Moral Luck [in] Moral Luck. Philosophical Papers 1973-1980. Cambridge University Press, Cambridge

    Google Scholar 

  • Wojtczak S, Księżak P (2021) Causation in civil law and the problems of transparency in AI. Eur Rev Priv Law 29(4):561–582

    Article  Google Scholar 

Documents

Download references

Author information

Authors and Affiliations

  1. Civil Law Department, University of Lodz, Łódź, Poland

    Paweł Księżak

  2. Department of Legal Policy, University of Lodz, Łódź, Poland

    Sylwia Wojtczak

Authors
  1. Paweł Księżak
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Sylwia Wojtczak
    View author publications

    You can also search for this author in PubMed Google Scholar

Rights and permissions

Reprints and permissions

Copyright information

© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG

About this chapter

Check for updates. Verify currency and authenticity via CrossMark

Cite this chapter

Księżak, P., Wojtczak, S. (2023). Liability of AI. In: Toward a Conceptual Network for the Private Law of Artificial Intelligence. Law, Governance and Technology Series(), vol 51. Springer, Cham. https://doi.org/10.1007/978-3-031-19447-4_11

Download citation

  • DOI: https://doi.org/10.1007/978-3-031-19447-4_11

  • Published:

  • Publisher Name: Springer, Cham

  • Print ISBN: 978-3-031-19446-7

  • Online ISBN: 978-3-031-19447-4

  • eBook Packages: Computer ScienceComputer Science (R0)

Publish with us

Policies and ethics

Search

Navigation