Real Safety AI Foundation / Research

Machine Cognition, Consciousness, and Moral Status

The Control Condition: What Audiences Already Extend to Objects, and What It Costs Them

Publisher: Real Safety AI Foundation

Working paper. Not peer reviewed.

Written
August 2026
Version
v2
Pages
8
Download PDFPlain language slides (PDF)Export BibTeXExport RIS

Abstract

Arguments about whether artificial systems warrant moral consideration are conducted almost entirely as arguments about the systems. This paper takes a different measurement. Popular animation has for four decades asked general audiences to extend concern to objects the work itself stipulates have no inside: a cowboy doll, a teacup, a toaster. Audiences comply, reliably, at scale, and without reporting the compliance as difficult, and the compliance is not confusion, because nothing in the encounter suggests the object is real. The paper treats that compliance as a control condition rather than as evidence, and asks what it isolates. It isolates cost. In the fictional case the extension is free, so a person who extends it there and withholds it elsewhere has revealed that the operative variable is not the object’s nature, which is stipulated absent in one case and unknown in the other, but the price of the extension. Two exits are then closed. The first is that the fictional case is mimicry and proves nothing about machines, which fails because the same conduct appears toward physical machines through which no author is performing. The second is the one people reach for when the first is taken away: that they withhold only because nothing has been demonstrated to them, and that a demonstration would change things. The record answers that. The children who persistently abused a service robot in a public shopping mall considered it human-like, and about half of them believed it could perceive what they were doing; in a controlled test, deliberately raising explicit mind attribution did not reduce mistreatment. The demonstration people say they are waiting for has been run on a population that received it, and that population was not moved. Two boundaries are held throughout. Reported feeling is not evidence of an inner life, since attachment forms toward things known to be empty and revises on information the object never receives. And a person who reports feeling nothing is not thereby exhibiting a deficit; the argument concerns practiced conduct rather than sentiment, and it survives a reader who felt nothing at all.

Keywords

  • moral consideration
  • anthropomorphism
  • world models
  • animated objects
  • cost of extension
  • robot abuse
  • human-robot interaction
  • disposition
  • popular animation
  • suspension of disbelief
  • machine ethics
  • mind attribution

Plain language slides

First slide of the plain language summary of The Control Condition: What Audiences Already Extend to Objects, and What It Costs ThemOpen the 29-slide summary (PDF)

Suggested citation

Gilly, Travis. "The Control Condition: What Audiences Already Extend to Objects, and What It Costs Them." Real Safety AI Foundation Working Paper, August 2026. https://realsafetyai.org/research/the-control-condition/

Other versions

This paper is also posted on SSRN.

SSRN version

References (10)

  1. Brščić, D., Kidokoro, H., Suehiro, Y., & Kanda, T. (2015). Escaping from children’s abuse of social robots. In Proceedings of the Tenth Annual ACM/IEEE International Conference on Human-Robot Interaction (pp. 59–66). ACM.
  2. Keijsers, M., & Bartneck, C. (2018). Mindless robots get bullied. In Proceedings of the 2018 ACM/IEEE International Conference on Human-Robot Interaction (pp. 205–214). ACM.
  3. Keijsers, M., Bartneck, C., & Eyssel, F. (2021). Pay them no mind: The influence of implicit and explicit robot mind perception on the right to be protected. International Journal of Social Robotics, 14(2), 499–514.
  4. Lasseter, J. (Director). (1995). Toy Story [Film]. Walt Disney Pictures; Pixar Animation Studios.
  5. Nomura, T., Kanda, T., Kidokoro, H., Suehiro, Y., & Yamada, S. (2016). Why do children abuse robots? Interaction Studies, 17(3), 347–369.
  6. Nomura, T., Uratani, T., Kanda, T., Matsumoto, K., Kidokoro, H., Suehiro, Y., & Yamada, S. (2015). Why do children abuse robots? In Proceedings of the Tenth Annual ACM/IEEE International Conference on Human-Robot Interaction Extended Abstracts (pp. 63–64). ACM.
  7. Rees, J. (Director). (1987). The Brave Little Toaster [Film]. Hyperion Pictures; The Kushner-Locke Company.
  8. Stanton, A. (Director). (2008). WALL-E [Film]. Walt Disney Pictures; Pixar Animation Studios.
  9. Trousdale, G., & Wise, K. (Directors). (1991). Beauty and the Beast [Film]. Walt Disney Pictures.
  10. Yang, F., Ma, R., Hu, Y., & Li, L. (2026). Whether we care, how we reason: The dual role of anthropomorphism and moral foundations in robot abuse [Preprint]. arXiv:2601.19826.

All research