Skip to main content
Log in

Predicting the Long-Term Effects of Human-Robot Interaction: A Reflection on Responsibility in Medical Robotics

  • Original Paper
  • Published:
Science and Engineering Ethics Aims and scope Submit manuscript

Abstract

This article addresses prospective and retrospective responsibility issues connected with medical robotics. It will be suggested that extant conceptual and legal frameworks are sufficient to address and properly settle most retrospective responsibility problems arising in connection with injuries caused by robot behaviours (which will be exemplified here by reference to harms occurred in surgical interventions supported by the Da Vinci robot, reported in the scientific literature and in the press). In addition, it will be pointed out that many prospective responsibility issues connected with medical robotics are nothing but well-known robotics engineering problems in disguise, which are routinely addressed by roboticists as part of their research and development activities: for this reason they do not raise particularly novel ethical issues. In contrast with this, it will be pointed out that novel and challenging prospective responsibility issues may emerge in connection with harmful events caused by normal robot behaviours. This point will be illustrated here in connection with the rehabilitation robot Lokomat.

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

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

Notes

  1. It is worth stressing that prospective responsibility problems—as they have been introduced here—concern the decision on whether it is ethically acceptable to deploy some robot (or robot technology) for a particular task, and not the decision on who will be (legally) responsible for harms caused by the robot: the two classes of problems are conceptually distinct. However, there may be connections between them. For example, one may argue that it is ethically acceptable to deploy some particular robot for some medical task provided that some particular rule for assigning legal responsibility in case of harms is imposed. A further analysis of these connections would be surely helpful to articulate and refine the claims concerning prospective responsibility which are introduced in this article.

  2. “It must be emphasized that the goal of introducing rehabilitation robots into clinics is not to replace physical and occupational therapists, but rather robots are a complement to existing treatment options” (Hidler et al. 2005, p. 22). Nevertheless, it is reasonable to believe that the reduction of health care costs is at least one of the main motives driving research in medical robotics; this is evident, for example, in Wasen’s (2010) reflections on robot-assisted surgery.

  3. One of the main techniques used to understand the neural process correlated with motor recovery in rehabilitation involves the recording of motor-evoked potentials during transcranial magnetic stimulation (TMS). In TMS, one applies a train of magnetic pulses on the cortex through an instrument placed on the scalp (Barker 1985; Barker 1999); the pulses induce an electric flow in the adjacent biological tissue by depolarizing nerve axons. To study brain plasticity, one applies a TMS onto a particular area of the motor cortex and observes the motor-evoked potential at the muscles of the subject: in this way one can detect correlations between stimulated brain areas and muscle activation, thus measure the size of brain areas that control the behaviour of individual muscles. There is some evidence (see for example Koski et al. 2004) that, in active rehabilitation, motor recovery is concurrent with increases in the size of motor areas (which would suggest recruitment of intact circuits to control impaired muscles) and with increases in motor excitability. See Van Der Loos and Reinkensmeyer (2008) for a discussion and references on the use of the Lokomat as a scientific tool to understand the mechanisms underlying motor rehabilitation.

  4. Machine learning technologies could be applied to let the robot learn to imitate the movements of human rehabilitation therapists. In this ideal scenario, the Lokomat undergoes a training step in which a human therapist moves the exoskeleton together with the patient’s lower limbs, while the system—possibly via neural networks—tries to learn the joints’ motor patterns so as to be able to reproduce them later. To reflect on this technological opportunity—which is undoubtedly interesting and stimulating—it is worth stressing that current machine learning technologies are strongly biased by some (often implicit) background assumptions on what is the “right” behaviour to learn (Datteri et al. 2005). Thus, in the ideal scenario envisaged here, the Lokomat would not learn from the human therapist’s movements alone, but the results of the learning process would be unavoidably dependent on background assumptions on how to perform “good” rehabilitation therapies. Scientific theories on neuro-rehabilitation cannot be easily dispensed with.

  5. The authors note that the clinical implications of the abnormal muscular patterns and of the pelvic restrictions are not necessarily negative: “… during early neurorehabilitation, it is often helpful if not mandatory to reduce the degrees of freedom through which a person can move since neurologically impaired patients can easily become overwhelmed with the amount of tasks to perform” (Hidler and Wall 2005, p. 191).

  6. Analogous considerations on the neurophysiological effects of brain-machines interfaces are discussed in Salvini et al. (2007). One should be careful to note that these provisional conclusions concern Lokomat-based therapies and not the Lokomat itself. Apart from the structural drawbacks discussed before (including the restrictions imposed by the robot to pelvic movements), the Lokomat is indeed a state-of-the-art, fully programmable and flexible robot. The concerns raised and discussed here regard essentially the way to program the robot so that it can deliver proper rehabilitation therapies, thus matching or even exceeding the efficiency of human-based therapies.

  7. FAQs of the Intuitive Surgical, available at http://www.intuitivesurgical.com/products/products_faq.html.

  8. Note that—as shown by the Da Vinci malfunctions discussed so far—behavioural anomalies may occur in highly tele-operated robots as well. Indeed, a joint bolt getting looser or a programming bug may result in an anomalous behaviour and cause injuries to the patient even if the robot is tele-operated—these faults can disrupt the strict correspondence between human and robot movements which is distinctive of tele-operated robots.

  9. http://www.paed.uscourts.gov/documents/opinions/09D0273P.pdf.

  10. http://www.ca3.uscourts.gov/opinarch/092042np.pdf.

  11. Dealing with these thorny philosophical and conceptual problems would be out of the scope of this article. See Woodward (2003) for a recent philosophical analysis of causality based on the notion of ‘intervention’ which may provide useful insights to understand and test causal ascriptions in human-robot interactions.

  12. A deeper analysis of the criteria for ascribing moral and legal responsibility in human-robot interaction scenarios is out of the scope of this paper. For a clear and insightful analysis of the role of the mythical image of robots in the Western countries with respect to responsibility ascription problems, see Mudry et al. (2008).

  13. This article has been chiefly concerned with robotic medical systems; an ethically-driven analysis of the potentialities and limitations of computer-only systems in medicine would be surely interesting and might reveal issues which (at least partly) overlap with those addressed here. To the present discussion on negligence and poor training it is worth noting that, often, computer simulations are used to train operators of robotic systems (cases in point are training of military drones through computer flight simulators). A computer simulator to be used for Da Vinci training must simulate not only the Da Vinci kinematics, but also the physical properties of the internal organs of a model patient the robot is interacting with. This is surely a technically challenging problem. Even though the research and development on Da Vinci simulators is still in its infancy, some products are commercially available and may help reduce the number of hours needed for achieving good skills (Lerner et al. 2010). Monitoring the developments of these technologies in the near future is surely important to articulate and sharpen the ethically-motivated reflections on negligence presented here.

References

  • Andonian, S., Okeke, Z., Okeke, D. A., Rastinehad, A., Vanderbrink, B. A., Richstone, L., et al. (2008). Device failures associated with patient injuries during robot-assisted laparoscopic surgeries: A comprehensive review of FDA MAUDE database. The Canadian Journal of Urology, 15(1), 3912–3916.

    Google Scholar 

  • Barker, A. T. (1985). Non-invasive magnetic stimulation of human motor cortex. The Lancet, 325(8437), 1106–1107. doi:10.1016/S0140-6736(85)92413-4.

    Article  Google Scholar 

  • Barker, A. T. (1999). The history and basic principles of magnetic nerve stimulation. Electroencephalography and Clinical Neurophysiology. Supplement, 51, 3–21.

    Google Scholar 

  • Borden, L. S., Kozlowski, P. M., Porter, C. R., & Corman, J. M. (2007). Mechanical failure rate of da Vinci robotic system. The Canadian Journal of Urology, 14(2), 3499–3501.

    Google Scholar 

  • Brink, G. (2002). Robotic surgery called a godsend. St. Petersburg Times. URL: http://www.sptimes.com/2002/10/31/TampaBay/Robotic_surgery_calle.shtml.

  • Carreyrou, J. (2010a). Botched operation using Da Vinci robot spurs lawsuit. The Wall Street Journal. URL: http://online.wsj.com/article/SB10001424052748703341904575266952674277806.html.

  • Carreyrou, J. (2010b). Surgical robot examined in injuries. The Wall Street Journal. URL: http://online.wsj.com/article/SB10001424052702304703104575173952145907526.html.

  • Coleman, J., & Mendlow, G. (2003). Theories of tort law. In E. Zalta (Ed.), Stanford Encyclopedia of Philosophy (2003rd ed.). Stanford: Center for the Study of Language and Information, Stanford University.

  • Colombo, G., Joerg, M., Schreier, R., & Dietz, V. (2000). Treadmill training of paraplegic patients using a robotic orthosis. Journal of Rehabilitation Research and Development, 37(6), 693–700.

    Google Scholar 

  • Datteri, E., Hosni, H., & Tamburrini, G. (2005). Machine learning from examples: A non-inductivist analysis. Logic & Philosophy of Science, 3(1), 1–31.

    Google Scholar 

  • Datteri, E., & Tamburrini, G. (2007). Biorobotic experiments for the discovery of biological mechanisms. Philosophy of Science, 74(3), 409–430

    Google Scholar 

  • Datteri, E., & Tamburrini, G. (2009). Ethical reflections on health care robotics. In R. Capurro & M. Nagenborg (Eds.), Ethics and Robotics (pp. 35–48). Amsterdam-Heidelberg: IOS Press/AKA.

    Google Scholar 

  • Eshleman, A. (2003). Moral responsibility. In E. Zalta (Ed.), Stanford Encyclopedia of Philosophy (2003rd ed.). Stanford: Center for the Study of Language and Information, Stanford University.

  • Hidler, J., Nichols, D., Pelliccio, M., & Brady, K. (2005). Advances in the understanding and treatment of stroke impairment using robotic devices. Topics in Stroke Rehabilitation, 12(2), 22–35.

    Article  Google Scholar 

  • Hidler, J., Nichols, D., Pelliccio, M., Brady, K., Campbell, D. D., Kahn, J. H., et al. (2009). Multicenter randomized clinical trial evaluating the effectiveness of the Lokomat in subacute stroke. Neurorehabilitation and Neural Repair, 23(1), 5–13. doi:10.1177/1545968308326632.

    Article  Google Scholar 

  • Hidler, J. M., & Wall, A. E. (2005). Alterations in muscle activation patterns during robotic-assisted walking. Clinical Biomechanics, 20(2), 184–193. doi:10.1016/j.clinbiomech.2004.09.016.

    Google Scholar 

  • Hornby, T. G., Campbell, D. D., Kahn, J. H., Demott, T., Moore, J. L., & Roth, H. R. (2008). Enhanced gait-related improvements after therapist- versus robotic-assisted locomotor training in subjects with chronic stroke: a randomized controlled study. Stroke: A Journal of Cerebral Circulation, 39(6), 1786–1792. doi:10.1161/STROKEAHA.107.504779.

    Article  Google Scholar 

  • Koliakos, N., Denaeyer, G., Willemsen, P., Schatteman, P., & Mottrie, A. (2008). Failure of a robotic arm during da Vinci prostatectomy: a case report. Journal of Robotic Surgery, 2(2), 95–96. doi:10.1007/s11701-008-0080-z (Springer, London).

    Google Scholar 

  • Koski, L., Mernar, T. J., & Dobkin, B. H. (2004). Immediate and long-term changes in corticomotor output in response to rehabilitation: correlation with functional improvements in chronic stroke. Neurorehabilitation and Neural Repair, 18(4), 230–249. doi:10.1177/1545968304269210.

    Article  Google Scholar 

  • Lerner, M. A., Ayalew, M., Peine, W. J., & Sundaram, C. P. (2010). Does training on a virtual reality robotic simulator improve performance on the Da Vinci surgical system? Journal of Endourology/Endourological Society, 24(3), 467–472. doi:10.1089/end.2009.0190.

    Article  Google Scholar 

  • Marchal-Crespo, L., & Reinkensmeyer, D. J. (2009). Review of control strategies for robotic movement training after neurologic injury. Journal of Neuroengineering and Rehabilitation, 6, 20. doi:10.1186/1743-0003-6-20.

    Article  Google Scholar 

  • Mudry, P.-A., Degallier, S., & Billard, A. (2008). On the influence of symbols and myths in the responsibility ascription problem in roboethics—A roboticist’s perspective. In Proceedings of RO-MAN 2008The 17th IEEE International Symposium on Robot and Human Interactive Communication (pp. 563–568). IEEE. doi:10.1109/ROMAN.2008.4600726.

  • Park, S. Y., Ahn, J. J.-K., Jeong, W., Ham, W. S., & Rha, K. H. (2010). A unique instrumental malfunction during robotic prostatectomy. Yonsei Medical Journal, 51(1), 148–150. doi:10.3349/ymj.2010.51.1.148.

    Article  Google Scholar 

  • Prassler, E., & Kosuge, K. (2008). Domestic robotics. In B. Siciliano & O. Khatib (Eds.), Springer Handbook of Robotics (pp. 1253–1282). New York: Springer.

    Chapter  Google Scholar 

  • Regnaux, J.-P., Saremi, K., Marehbian, J., Bussel, B., & Dobkin, B. H. (2008). An accelerometry-based comparison of 2 robotic assistive devices for treadmill training of gait. Neurorehabilitation and Neural Repair, 22(4), 348–354. doi:10.1177/1545968307310050.

    Article  Google Scholar 

  • Salvini, P., Datteri, E., Laschi, C., & Dario, P. (2007). Scientific models and ethical issues in hybrid bionic systems research. AI & Society, 22(3), 431–448. doi:10.1007/s00146-007-0158-6 (Springer, London).

  • Santoro, M., Marino, D., & Tamburrini, G. (2007). Learning robots interacting with humans: From epistemic risk to responsibility. Ai & Society, 22(3), 301–314. doi:10.1007/s00146-007-0155-9.

    Article  Google Scholar 

  • Shepherd, R. B. (2001). Exercise and training to optimize functional motor performance in stroke: Driving neural reorganization? Neural Plasticity, 8(1–2), 121–129. doi:10.1155/NP.2001.121.

    Article  Google Scholar 

  • Sinkaer, T., & Popović, D. B. (2009). Neurorehabilitation technologies—Present and future possibilities. Introduction. NeuroRehabilitation, 25(1), 1–3. doi:10.3233/NRE-2009-0494.

  • Swinnen, E., Duerinck, S., Baeyens, J.-P., Meeusen, R., & Kerckhofs, E. (2010). Effectiveness of robot-assisted gait training in persons with spinal cord injury: A systematic review. Journal of Rehabilitation Medicine, 42(6), 520–526. doi:10.2340/16501977-0538.

    Article  Google Scholar 

  • Tamburrini, G. (2009). Robot ethics: A view from the philosophy of science. In R. Capurro & M. Nagemborg (Eds.), Ethics and Robotics (pp. 11–22). Amsterdam-Heidelberg: IOS Press/AKA.

    Google Scholar 

  • Tamburrini, G., & Datteri, E. (2005). Machine experiments and theoretical modelling: From cybernetic methodology to neuro-robotics. Minds and Machines, 15(3–4), 335–358. doi:10.1007/s11023-005-2924-x.

    Article  Google Scholar 

  • Taylor, R. H., Menciassi, A., Fichtinger, G., & Dario, P. (2008). Medical robotics and computer-integrated surgery. In B. Siciliano & O. Khatib (Eds.), Springer Handbook of Robotics (pp. 1199–1222). Springer, New York. doi:10.1007/978-3-540-30301-5_53.

  • Van Der Loos, H. F. M., & Reinkensmeyer, D. J. (2008). Rehabilitation and health care robotics. In B. Siciliano & O. Khatib (Eds.), Springer Handbook of Robotics (pp. 1223–1251). Berlin, Heidelberg: Springer. doi:10.1007/978-3-540-30301-5.

  • Wasen, K. (2010). Replacement of highly educated surgical assistants by robot technology in working life: Paradigm shift in the service sector. International Journal of Social Robotics, 2(4), 431–438. doi:10.1007/s12369-010-0062-y.

    Article  Google Scholar 

  • Woodward, J. (2003). Making Things Happen: A Theory of Causal Explanation. USA: Oxford University Press.

    Google Scholar 

  • Zorn, K. C., Gofrit, O. N., Orvieto, M. A., Mikhail, A. A., Galocy, R. M., Shalhav, A. L., et al. (2007). Da Vinci robot error and failure rates: single institution experience on a single three-arm robot unit of more than 700 consecutive robot-assisted laparoscopic radical prostatectomies. Journal of Endourology/Endourological Society, 21(11), 1341–1344. doi:10.1089/end.2006.0455.

    Article  Google Scholar 

Download references

Acknowledgments

I am grateful to Guglielmo Tamburrini for stimulating discussions on the topics addressed here, and two anonymous referees for their valuable comments and criticisms on earlier versions of this paper.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Edoardo Datteri.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Datteri, E. Predicting the Long-Term Effects of Human-Robot Interaction: A Reflection on Responsibility in Medical Robotics. Sci Eng Ethics 19, 139–160 (2013). https://doi.org/10.1007/s11948-011-9301-3

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11948-011-9301-3

Keywords

Navigation