[Paper Review] Collecting the Public Perception of AI and Robot Rights

Gabriel Lima, Changyeon Kim+3 more|arXiv (Cornell University)|Aug 4, 2020

Psychology of Moral and Emotional JudgmentNeuroscience41 references46 citations

TL;DR

An experiment with 1270 online participants (plus a 164-replication) surveys 11 potential rights for AI/robots and tests whether misperception-debunking interventions shift attitudes, finding general opposition but a boost for the right against cruel treatment and notable stance changes from interventions.

ABSTRACT

Whether to give rights to artificial intelligence (AI) and robots has been a sensitive topic since the European Parliament proposed advanced robots could be granted "electronic personalities." Numerous scholars who favor or disfavor its feasibility have participated in the debate. This paper presents an experiment (N=1270) that 1) collects online users' first impressions of 11 possible rights that could be granted to autonomous electronic agents of the future and 2) examines whether debunking common misconceptions on the proposal modifies one's stance toward the issue. The results indicate that even though online users mainly disfavor AI and robot rights, they are supportive of protecting electronic agents from cruelty (i.e., favor the right against cruel treatment). Furthermore, people's perceptions became more positive when given information about rights-bearing non-human entities or myth-refuting statements. The style used to introduce AI and robot rights significantly affected how the participants perceived the proposal, similar to the way metaphors function in creating laws. For robustness, we repeated the experiment over a more representative sample of U.S. residents (N=164) and found that perceptions gathered from online users and those by the general population are similar.

Motivation & Objective

• Assess online public perceptions of 11 potential rights for future autonomous AI and robots.
• Test whether interventions that debunk misconceptions alter people’s stance on AI/robot rights.
• Examine the influence of presentation style, agent type, and autonomy level on perceived rights.
• Compare online MTurk results with a demographically representative US sample to assess robustness.

Proposed method

• Survey design presenting 11 rights as declarative statements to participants with randomized phrasing styles (Indirect, Direct, Explicit).
• Two agent forms (AI/robot with/without physical body) and two autonomy levels (fully autonomous vs not) were used.
• Four intervention designs (normative, descriptive, non-human, human) each using 6–9 visual cards to debunk misconceptions.
• Attention checks and time controls ensured engagement; responses were cleaned to yield N=1270 valid responses from MTurk (US).
• A demographically representative US sample (N=164) was collected to test cross-sample robustness.
• Outcomes included pre-stance and post-stance on 11 rights, stance-change, and effect-size statistics (Hedge’s g).

Experimental results

Research questions

• RQ1 What are online public perceptions of 11 rights that might apply to future AI and robots?
• RQ2 Can interventions debunking common misconceptions meaningfully shift these perceptions?
• RQ3 How do agent type, autonomy level, and phrasing style influence stance changes?
• RQ4 Are perceptions among online (MTurk) participants consistent with a demographically representative sample?

Key findings

• Online participants generally opposed granting most rights to AI and robots, with the exception of support for the right against cruel punishment and treatment.
• Non-human intervention produced the largest stance-change across rights (g = 0.440; μ_sc = 0.408).
• Contracts, the right to sue/be sued, and the right to hold assets showed the largest stance changes under the non-human design (g = 0.442, μ_sc = 0.582; g = 0.439, μ_sc = 0.598; g = 0.421, μ_sc = 0.546 respectively).
• Explicitly presenting agents as fully autonomous slightly increased stance-change (g = 0.134).
• Longer intervention reading time correlated with greater stance-change; higher initial NARS scores correlated with larger changes.
• The replication with a representative US sample (N=164) yielded similar trend patterns to the MTurk online sample.