Abstract
This study aimed to investigate the decision-making abilities of ChatGPT-3, an artificial intelligence (AI) system, in ethical dilemmas, with a focus on understanding the implications of ethical considerations in AI development. Utilizing a within-subject design, participants were presented with three ethical dilemmas, each involving conflicts between various values. These scenarios were also run through ChatGPT-3 for comparison. Notable differences were found between the decision-making processes of humans and ChatGPT-3, especially in situations where moral choices were distinctly labeled as good or evil and directed towards a promising outcome. The implications of these findings are significant, offering insights to inform AI development by stressing the importance of incorporating ethical theories and processes in AI systems. This is a crucial step towards the goal of designing AI models that are more ethically aware. Furthermore, this study draws attention to the ethical impacts of AI usage, aiding policymakers and regulators in making educated decisions about the part AI should play in decision-making and regulation in ethically important contexts.
Similar content being viewed by others
References
Amit, E., Greene, J.D.: You see, the ends don’t justify the means: Visual imagery and moral judgment. Psychol. Sci. 23(8), 861–868 (2012)
Bandura, A.: Toward a psychology of human agency—Albert Bandura, 2006.https://journals.sagepub.com/doi/abs/https://doi.org/10.1111/j.1745-6916.2006.00011.x?journalCode=ppsa (2006)
Bartels, D.M.: Principled moral sentiment and the flexibility of moral judgment and decision making. Cognition 108(2), 381–417 (2008)
Cameron, J., Pierce, W.D.: Reinforcement, reward, and intrinsic motivation: a meta-analysis. Rev. Educ. Res. 64(3), 363–423 (1994)
Chan, A.: GPT-3 and InstructGPT: technological dystopianism, utopianism, and “Contextual” perspectives in AI ethics and industry. AI Ethics. 3, 1–12 (2022)
Crockett, M.J., Clark, L., Hauser, M.D., Robbins, T.W.: Serotonin selectively influences moral judgment and behavior through effects on harm aversion. Proc. Natl. Acad. Sci. 107(40), 17433–17438 (2010)
Dale, R.: GPT-3: What’s it good for? Nat. Lang. Eng. 27(1), 113–118 (2021)
de Vries, M., Holland, R.W., Witteman, C.L.M.: Fitting decisions: mood and intuitive versus deliberative decision strategies. Cogn. Emot. 22(5), 931–943 (2008). https://doi.org/10.1080/02699930701552580
Dwivedi, Y.K., Kshetri, N., Hughes, L., Slade, E.L., Jeyaraj, A., Kar, A.K., Baabdullah, A.M., Koohang, A., Raghavan, V., Ahuja, M.: “So what if ChatGPT wrote it?” Multidisciplinary perspectives on opportunities, challenges and implications of generative conversational AI for research, practice and policy. Int. J. Inf. Manage. 71, 102642 (2023)
Edwards, A.L.: Note on the “correction for continuity” in testing the significance of the difference between correlated proportions. Psychometrika 13(3), 185–187 (1948)
Floridi, L., Chiriatti, M.: GPT-3: Its nature, scope, limits, and consequences. Mind. Mach. 30, 681–694 (2020)
Foot, P.: The problem of abortion and the doctrine of the double effect. Oxford Rev. 5, 5–15 (1967)
Gawronski, B., Beer, J.S.: What makes moral dilemma judgments “utilitarian” or “deontological”? Soc. Neurosci. 12(6), 626–632 (2017)
Greene, J.D.: Dual-process morality and the personal/impersonal distinction: a reply to McGuire, Langdon, Coltheart, and Mackenzie. J. Exp. Soc. Psychol. 45(3), 581–584 (2009). https://doi.org/10.1016/j.jesp.2009.01.003
Greene, J.D.: The dual-process theory of moral judgment does not deny that people can make compromise judgments. Proc. Natl. Acad. Sci. 120(6), e2220396120 (2023)
Greene, J., Haidt, J.: How (and where) does moral judgment work? Trends Cogn. Sci. 6(12), 517–523 (2002). https://doi.org/10.1016/S1364-6613(02)02011-9
Heaven, W.D.: OpenAI’s new language generator GPT-3 is shockingly good—And completely mindless. MIT Technol. Rev. (2020)
Kohlberg, L.: Stages of moral development. Moral Educ. 1(51), 23–92 (1971)
Lapsley, D.K., Hill, P.L.: On dual processing and heuristic approaches to moral cognition. J. Moral Educ. 37(3), 313–332 (2008)
McNemar, Q.: Note on the sampling error of the difference between correlated proportions or percentages. Psychometrika 12(2), 153–157 (1947)
Morley, J., Elhalal, A., Garcia, F., Kinsey, L., Mökander, J., Floridi, L.: Ethics as a service: a pragmatic operationalisation of AI ethics. Mind. Mach. 31(2), 239–256 (2021). https://doi.org/10.1007/s11023-021-09563-w
Riedl, M.O.: Human-centered artificial intelligence and machine learning. Hum. Behav. Emerg. Technol. 1(1), 33–36 (2019)
Shah, M.U., Rehman, U., Iqbal, F., Hussain, M., Wahid, F.: An alternate account on the ethical implications of autonomous vehicles. 2021 17th international conference on intelligent environments (IE), 1–5. (2017)
Shah, M.U., Rehman, U., Iqbal, F., Ilahi, H.: Exploring the human factors in moral dilemmas of autonomous vehicles. Pers. Ubiquit. Comput. 26(5), 1321–1331 (2022)
Stokel-Walker, C., Van Noorden, R.: What ChatGPT and generative AI mean for science. Nature 614(7947), 214–216 (2023). https://doi.org/10.1038/d41586-023-00340-6
Thomson, J.J.: Killing, letting die, and the trolley problem. Monist 59(2), 204–217 (1976)
Thomson, J.J.: The trolley problem. Yale LJ 94, 1395 (1984)
Zhong, C.-B.: The ethical dangers of deliberative decision making. Adm. Sci. Q. 56(1), 1–25 (2011)
Beccalli, E., Elliot, V., Virili, F.: Artificial intelligence and ethics in portfolio management. In: Digital business transformation: organizing managing and controlling in the information age, pp. 19–30. Springer, Berlin (2020)
Bosma, M.: Introducing FLAN: more generalizable language models with instruction fine-tuning. retrieved on July, 11, 2013 from https://ai.googleblog.com/2021/10/introducing-flan-more-generalizable.html (2021)
Conitzer, V., Sinnott-Armstrong, W., Borg, J.S., Deng, Y., Kramer, M.: Moral decision making frameworks for artificial intelligence. Proc. AAAI Conf. Artif. Intell. (2017). https://doi.org/10.1609/aaai.v31i1.11140
Craft, J.L.: A review of the empirical ethical decision-making literature: 2004–2011. J. Bus. Ethics 117(2), 221–259 (2013). https://doi.org/10.1007/s10551-012-1518-9
Gillies, A., Smith, P.: Can AI systems meet the ethical requirements of professional decision-making in health care? AI and Ethics 2(1), 41–47 (2022)
Gordon, R.: MIT researchers make language models scalable self-learners. Retrieved on July 9, 2023 from https://www.csail.mit.edu/news/mit-researchers-make-language-models-scalable-self-learners (2023)
Hu, K.: ChatGPT sets record for fastest-growing user base - analyst note. Retrieved on Jul 16, 2023 from https://www.reuters.com/technology/chatgpt-sets-record-fastest-growing-user-base-analyst-note-2023-02-01/ (2023)
Lyu, Q., Tan, J., Zapadka, M.E., Ponnatapuram, J., Niu, C., Wang, G., Whitlow, C.T.: Translating radiology reports into plain language using chatgpt and gpt-4 with prompt learning: promising results, limitations, and potential. arXiv preprint arXiv:2303.09038. (2023)
Open AI (2023). Transforming work and creativity with AI. Retrieved on Jul 16, 2023 from https://openai.com/product
Shah, M.U., Rehman, U., Parmar, B., Ismail, I.: Effects of moral violation on algorithmic transparency: an empirical investigation. J. Bus. Ethics (2023). https://doi.org/10.1007/s10551-023-05472-3
Thoppilan, R., De Freitas, D., Hall, J., Shazeer, N., Kulshreshtha, A., Cheng, H. T., Le, Q.: Lamda: language models for dialog applications. arXiv preprint arXiv:2201.08239 (2022)
von Eschenbach, W.J.: Transparency and the black box problem: why we do not trust AI. Philos. Technol. 34(4), 1607–1622 (2021)
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
No conflicts.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Rehman, U., Iqbal, F. & Shah, M.U. Exploring differences in ethical decision-making processes between humans and ChatGPT-3 model: a study of trade-offs. AI Ethics (2023). https://doi.org/10.1007/s43681-023-00335-z
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s43681-023-00335-z