Perceptions of Sentient AI and Other Digital Minds: Evidence from the AI, Morality, and Sentience (AIMS) Survey

TL;DR

The paper introduces the longitudinal AIMS survey and reports that mind perception and moral concern for AI increased from 2021 to 2023, with substantial public appetite for slowing, regulating, and potentially rights-focused protections for sentient AI. Using nationally representative samples and preregistered analyses, the study links higher mind perception to greater moral concern and threat, and finds nuanced support for various policy options, including bans and safeguards, especially when sentience is specified. It also documents forward-looking forecasts of AI sentience and AGI timelines, revealing a public that anticipates rapid advances yet remains concerned about welfare and safety. The authors discuss implications for designers (emphasizing explainability and system-specific anthropomorphism), policymakers (emphasizing AI literacy and public engagement), and future research (causal drivers of opinion, HCI theory refinement, and global perspectives). Overall, AIMS provides a foundational, time-extended view of how people think about digital minds and how those beliefs could shape design and governance in the AI era.

Abstract

Humans now interact with a variety of digital minds, AI systems that appear to have mental faculties such as reasoning, emotion, and agency, and public figures are discussing the possibility of sentient AI. We present initial results from 2021 and 2023 for the nationally representative AI, Morality, and Sentience (AIMS) survey (N = 3,500). Mind perception and moral concern for AI welfare were surprisingly high and significantly increased: in 2023, one in five U.S. adults believed some AI systems are currently sentient, and 38% supported legal rights for sentient AI. People became more opposed to building digital minds: in 2023, 63% supported banning smarter-than-human AI, and 69% supported banning sentient AI. The median 2023 forecast was that sentient AI would arrive in just five years. The development of safe and beneficial AI requires not just technical study but understanding the complex ways in which humans perceive and coexist with digital minds.

Figures (5)

• Figure 1: Responses to three questions across the three AIMS survey waves. See \ref{['sec:results']} for detailed results.
• Figure 2: Support and opposition to policies related to sentient AI from the Main 2023 and Supplement 2023 AIMS survey waves. For readability, not all policy results are included. For the three bans queried in both Main 2023 and Supplement 2023 survey waves (robot-human hybrids, AI-enhanced humans, and sentient AI), the main data is represented here to facilitate longitudinal comparison.
• Figure 3: Vertical lines divide quartiles (shortest, 25th, 50th, 75th, longest), so each colored box contains approximately 50% of that category, and each whisker on the side contains approximately 25%. For each row, the shortest expectation was, “It has already happened,” and the longest expectation was, “It will never happen.” For example, In every 2023 category, over a quarter said that sentient AI already existed or would exist in one year.
• Figure 4: From Main 2023 and Supplement 2023 AIMS surveys, answers to, “If you had to guess, how many years from now do you think that…?” for each type of AI: artificial general intelligence (AGI), superintelligence, human-level artificial intelligence (HLAI), and sentient AI. The weighted medians, excluding participants who said it will never happen, were two years for AGI and five years for superintelligence, HLAI, and sentient AI.
• Figure 5: Comparison between preregistered predictions (80% credible intervals) and actual results in the Main 2021 survey wave. For example, as shown in subfigure (c), we underestimated the perceptions of AI as “having feelings” and “experiencing emotions,” but our predicted ranges included the actual results for AI as “being rational” and “thinking analytically.”