The problem of imitation in human-AI emotional relations

PROJECT SETTLED 

Project title: The problem of imitation in human-AI emotional relations
Principal investigator: dr Emilia Kaczmarek
Financing: UW, IDUB POB-IV
Project Period: 2022-2023
Grant no. 501-D117-20-4004310

Project goals: Thanks to the development of artificial intelligence, social robots, and chatbots, there are more and more devices with which people can engage in emotional relationships. Currently, these devices are not capable of reciprocating human emotions; however, they can respond in a way that seems as if they harbor feelings toward their users. This phenomenon raises the question of the moral evaluation of emotional relationships between humans and artificial intelligence entities that give their users the illusion of reciprocity. The aim of the project is to examine whether these relationships are indeed deceptive in nature and whether this issue may justify restrictions on the development of technologies designed to evoke emotions in humans. Throughout the project, a thorough analysis of the “deception” arguments raised in the debate on the ethical aspects of emotional relationships between humans and artificial intelligence will be conducted.

Keywords: AI ethics, social chatbots, AI-girlfriend, AI-friends, social robots, self-deception, robotic deception


Publication of the project results

Papers:

Conference presentations:

  • Kaczmarek, E., The problem of deception in human-AI emotional relations at Harvard University during the Tech Ethics eXchange NorthEast (teχnē) scientific conference, organized by Social and Ethical Responsibilities of Computing at MIT, the Ethics Institute at Northeastern University, and Embedded EthiCS at Harvard University, on May 18–19, 2023.
  • Kaczmarek, E., Fighting loneliness with chatbots? Ethics of self-deception, The annual EACME Conference, University of Warsaw, 14–16 September 2023
  • Kaczmarek, E., Comparing the risks of under-attributing and over-attributing moral status to AI, AI-Phi-Rel 2024 Conference, University of Warsaw.

 

Popularization of results:

  • Kaczmarek, E., Idealny przyjaciel bot?, Kultura Liberalna, Nr 742 (12/2023)
  • Podcast, Cybe(r)ewolucja, Czy AI należą się prawa i jak sprawdzić, kiedy będzie świadoma?,
    Wyborcza.biz, 2024
  • Conducting a seminar at the Faculty of Philosophy, University of Warsaw, titled ‘Human-AI Emotional Interaction: Ethics of Self-Deception’ (2023)