Standard

IEEE WHITE PAPER

Published

Note: This standard has a new edition: IEEE WHITE PAPER

Singles purchase not accessible

This standard can not be purchased as a single sales.

Abstract

INTRODUCTION AND SCOPE Deception, defined here as misleading others into believing something that the deceiver does not believe in (DePaulo et al. 2003 [17]1), is often viewed negatively. This is because it involves misleading or manipulating others, which can erode trust, harm relationships, and cause ethical concerns. However, deception is not always considered inherently bad; its moral and ethical evaluation arguably depends on the context, intent, and consequences. Similarly, if one views deception through different philosophical prisms, one sees different outcomes. For example, a deontologist may say that as a rule, any deception in a human–empathic AI partner relationship is always bad; while a utilitarian will ask if a greater overall good is served by deception in this relationship; a virtue ethicist may conclude that it depends on the character and intentions behind the system; a pragmatist may say that the moral acceptability of deception depends on the context; and a Shintoist may be more open to the presence of a subject in the object, depending on the intent and curation of the system. The purpose of this white paper is to think through the parameters of debate about deception and empathy-based human-AI partnering looking through the prism of IEEE P7014.1 issues, but mindful that deception in relation to AI is a broader topic. Funded by a UK Responsible AI award2 and organized by Project AEGIS,3 it explicitly draws on academic thought from several disciplines to distil a range of ethical positions. This is done to advance the work of the IEEE P7014.1 working group, an IEEE effort to develop a recommended practice to define ethical considerations and good practices regarding the use of emulated empathy in GPAI systems for human-AI partnerships. The white paper does not address all possible issues raised by empathy-based human-AI partnerships, but instead focuses on the issue of deception. Similarly, the white paper solely represents the views of the authors and does not necessarily represent a position of either the IEEE P7014.1 working group, the IEEE SSIT Standards Committee, the IEEE, or the IEEE Standards Association. 1 The numbers in brackets correspond to those of the references in Section 6.

Document information

  • Standard from IEEE_AC
  • Published:
  • Version: 0
  • Document type: IS
  • Additional information
    • IEEE P7014.1