Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Views on AI aren't binary -- they're plural

Thorin Bristow, Luke Thorburn, Diana Acosta-Navas

TL;DR

The paper argues that the common split between AI Ethics and AI Safety is a false dichotomy and that the AI governance landscape is a plural, multi-axial space. By dissecting stereotypes (Ethics and Alignment) and their discontents, it demonstrates overlaps, tensions, and the existence of multiple camps beyond a binary division. It advocates bridge-building strategies, including holistic institutions, broad stakeholder collaboration, empirical testing of contested claims, and democratic governance to reduce polarization. The work highlights the risks of framing AI risk discussions in binary terms, which can be co-opted by corporate and political interests, and emphasizes a value-pluralist path toward more inclusive and effective responsible AI.

Abstract

Recent developments in AI have brought broader attention to tensions between two overlapping communities, "AI Ethics" and "AI Safety." In this article we (i) characterize this false binary, (ii) argue that a simple binary is not an accurate model of AI discourse, and (iii) provide concrete suggestions for how individuals can help avoid the emergence of us-vs-them conflict in the broad community of people working on AI development and governance. While we focus on "AI Ethics" an "AI Safety," the general lessons apply to related tensions, including those between accelerationist ("e/acc") and cautious stances on AI development.

Views on AI aren't binary -- they're plural

TL;DR

The paper argues that the common split between AI Ethics and AI Safety is a false dichotomy and that the AI governance landscape is a plural, multi-axial space. By dissecting stereotypes (Ethics and Alignment) and their discontents, it demonstrates overlaps, tensions, and the existence of multiple camps beyond a binary division. It advocates bridge-building strategies, including holistic institutions, broad stakeholder collaboration, empirical testing of contested claims, and democratic governance to reduce polarization. The work highlights the risks of framing AI risk discussions in binary terms, which can be co-opted by corporate and political interests, and emphasizes a value-pluralist path toward more inclusive and effective responsible AI.

Abstract

Recent developments in AI have brought broader attention to tensions between two overlapping communities, "AI Ethics" and "AI Safety." In this article we (i) characterize this false binary, (ii) argue that a simple binary is not an accurate model of AI discourse, and (iii) provide concrete suggestions for how individuals can help avoid the emergence of us-vs-them conflict in the broad community of people working on AI development and governance. While we focus on "AI Ethics" an "AI Safety," the general lessons apply to related tensions, including those between accelerationist ("e/acc") and cautious stances on AI development.
Paper Structure (16 sections, 3 figures)

This paper contains 16 sections, 3 figures.

Table of Contents

  1. The false binary
  2. The caricature
  3. A note on language
  4. Ethics (the stereotype)
  5. Alignment (the stereotype)
  6. Ethics's discontents with Alignment
  7. Alignment's discontents with Ethics
  8. The complex reality
  9. Where Ethics and Alignment (actually) differ
  10. Where Ethics and Alignment (actually) are similar
  11. Complication: There are more than two camps
  12. Complication: The existential risk narrative has corporate value
  13. Complication: Good and bad ideas can be knotted together
  14. Overcoming the dichotomy
  15. Why should we?
  16. ...and 1 more sections

Figures (3)

  • Figure 1: Examples of the many axes along which views on AI vary. Adapted from bluemke2023.
  • Figure 2: Attempts to map the space of perspectives on AI.
  • Figure 3: Another map, from bensinger2024.