Preventing Sociopathic Robots

Study Status

Modeling Design
Modeling Benchmarking
Manuscripts Preparation
Publication

Preventing Sociopathic Robots

Leveraging Vulnerability and Homeostasis to Encourage Prosocial Behavior in World-Modeling AI

We will run a variety of simulations to demonstrate how prosocially aligned AI may be emergently bootstrapped via vulnerable engagement (alone and in competition/collaboration with others) with an open-ended Minecraft-like virtual world.

Such environments are well-suited for training and testing for our systems, who have several planned emergent capacities: (I) Abilities for meta-learning and counterfactual modeling; (II) Vulnerability; (III) Incorporation of biologically-inspired inductive biases.

Combined with meta-learning and counterfactual modeling that can accommodate complex agent-based relationships across multiple time scales, we will demonstrate how social alignment may emerge as a kind of predictive homeostasis (i.e., allostasis, or “sociostasis”).

https://i0.wp.com/advancedconsciousness.org/wp-content/uploads/2023/05/ChristovMoore_A_vulnerable_homeostatic_Buddhist_robot_learning__f2545e91-bb30-4d7d-8101-52e4b4e70310.jpg?w=891

https://i0.wp.com/advancedconsciousness.org/wp-content/uploads/2023/05/ChristovMoore_A_vulnerable_homeostatic_robot_learning_to_experi_90767625-657d-41b0-b8d7-94604da7e623.jpg?w=891

In close collaboration with other groups dedicated to human-centered AI, we will focus on determining the extent to which human-mimetic design may allow for the development of effective and robustly beneficial artificial intelligences.

We hope to a) determine conditions that are more or less likely to result in either prosocial or antisocial ‘personalities’, b) evaluate the consequences of these agent traits on interpersonal interactions, and c) explore the degree to which these emergent social dynamics alter the personalities of our agents.

Outcomes

2023

Christov-Moore, L., Reggente, N., Vaccaro, A., Schoeller, F., Pluimer, B., Douglas, P. K., Iacoboni, M., Man, K., Damasio, A., & Kaplan, J. T. (2023). Preventing antisocial robots: A pathway to artificial empathy. Sci. Robot, 8, eabq3658. Preventing antisocial robots: A pathway to artificial empathy. [PDF] [Blog Writeup]
Safron, A., Sakthivadivel, D., Sheikhbahaee, Z., Bein, M., Razi, A., Levin, M. (2023, April). Making and Breaking Symmetries in Mind and Life. In “Making and Breaking Symmetries in Mind and Life” a special issue of Interface Focus, 13. Making and Breaking Symmetries in Mind and Life. [PDF]
Sheikhbahaee, Z., Safron, A., Hesp, C., & Dumas, G. (2023, December). From Physics to Sentience: Deciphering the Semantics of the Free-Energy Principle and Evaluating its Claims. Physics of Life Reviews, 47, 276 - 278. From Physics to Sentience: Deciphering the Semantics of the Free-Energy Principle and Evaluating its Claims. *Equal contributions* [PDF]
Christov‐Moore, L., Jinich‐Diamant, A., Safron, A., Lynch, C., & Reggente, N. (2023). Cognitive Science Below the Neck: Toward an Integrative Account of Consciousness in the Body. Cognitive Science, 47(3), e13264. Cognitive Science Below the Neck: Toward an Integrative Account of Consciousness in the Body. [PDF] [Blog Writeup]

2022

Christov-Moore, L., Reggente, N., Vaccaro, A., Schoeller, F., Pluimer, B., Douglas, P. K., Iacaboni, M., Kaplan, J. (2022). Are Robots Sociopaths? A Neuroscientific Approach to the Alignment Problem. Organization for Human Brain Mapping, Glasgow, Scotland. [Poster]

2023

Moore, C. (2023, October 25-26). Emulating Entheogens for Enhancing Empathy. Compassion 2.0, San Francisco, CA. [Talk] [Supplemental]

2024

- - Ubiquity University. (2024). Humanity Rising Day 935: IACS III: Aesthetic Chills and Empathic AI [YouTube]. https://www.youtube.com/watch?v=IV9uJKAvBKo
  - Ubiquity University. (2024). Humanity Rising Day 934: IACS II: Toward a multiscale account of trust [YouTube]. https://www.youtube.com/watch?v=GaUtEtU0otE

2023

- - Digital, O., & Lorenzo, A. D. (2023). IA boazinha: estudo sugere novo método para criar robôs com empatia. Olhar Digital. IA boazinha: estudo sugere novo método para criar robôs com empatia.
  - Greengard, S. (2023). Making Empathy Artificial. Communications of the ACM. Making Empathy Artificial.
  - Atomic Podcast with Dr. Rooney Sappington. (2023). Atomic Podcast EP. 14 - Interview with Dr. Nicco Reggente and Dr. Leonardo Christov-Moore [Video]. YouTube. Atomic Podcast EP. 14 - Interview with Dr. Nicco Reggente and Dr. Leonardo Christov-Moore [Video]
  - Crespi, S., Mclean, K., Lloreda, L.C. (2023). The AI special issue, adding empathy to robots, and scientists leaving Arecibo. Science. The AI special issue, adding empathy to robots, and scientists leaving Arecibo.
  - Fan, S. (2023). Giving AI a Sense of Empathy Could Protect Us From Its Worst Impulses. Singularity Hub. Giving AI a Sense of Empathy Could Protect Us From Its Worst Impulses.
  - Yirka. B. (2023). How to give AI-based robots empathy so they won't want to kill us. Tech Xplore. How to give AI-based robots empathy so they won't want to kill us.
  - (2023). How to give AI-based robots empathy so they won't want to kill us. News Azi. Giving AI a Sense of Empathy Could Protect Us From Its Worst Impulses.
  - (2023). How to give AI-based robots empathy so they won't want to kill us. Today Headline. How to give AI-based robots empathy so they won't want to kill us.