This research provided a novel experimental exploration of relationship style stigma (RSS) targeting consensually nonmonogamous (CNM) prospective parents, as a function of their sexual orientation. Across three preregistered experiments, Mechanical Turk participants (N = 1,353) rated descriptions of highly functioning partners allegedly written by adoption social workers. Targets were identical apart from relationship style (monogamous, cheating, open, or polyamorous; Experiments 1-3) and sexual orientation (heterosexual, lesbian, or gay; Experiment 3). Robust RSS emerged; relative to identical CNM targets, monogamous partners were rated as superior (e.g., inspired less social distance and were viewed as more moral and having better relationships) and were more frequently approved to adopt a child (who was expected to have superior outcomes). Across experiments, CNM partners were over 15 times more likely to be denied adoption than identical monogamous partners. CNM partners were typically evaluated more poorly than cheating partners, suggesting that it may be viewed as preferable to fail at monogamy than purposefully attempt nonmonogamy. RSS emerged regardless of targets' sexual orientation, though effects were somewhat attenuated for lesbians. Results were moderated by participants' political orientation, such that RSS was more robust among conservatives. Targets' perceived morality and relationship strength mediated RSS even when accounting for expected stigma targeting CNM partners. (PsycInfo Database Record (c) 2026 APA, all rights reserved).
Generative artificial intelligence (AI) is reshaping creative industries, raising questions about how people evaluate AI-generated work. Building on extant work, we synthesize what we term the mind-superiority hypothesis, which holds that entities perceived as possessing a "mind" receive more favorable evaluations, particularly in creative domains. We apply this framework to the novel context of AI-generated creative work. We first explore whether AI-generated creative work is judged less favorably than human-created work because generative AI is seen as having lower mental capabilities. We then examine whether mentalistic mindset prompts-instructions encouraging observers to interpret an entity in terms of mental activities such as thinking and feeling-can heighten perceived mental capabilities and, in turn, improve evaluations. We also identify reasoning disclosure as a practical intervention that enhances mind perception and boosts evaluations of AI's creative output. Across seven studies (N = 2,333), the results provide converging evidence supporting the mind-superiority hypothesis and provide a theoretically grounded account of why AI-generated creative work is often evaluated less favorably-and how such evaluations can be improved. (PsycInfo Database Record (c) 2026 APA, all rights reserved).
Many real-world, safety-critical domains require searching for targets in complex scenes, like a pilot scanning a runway for hazards. Augmented reality (AR) displays show promise for faster, more accurate attention guidance during visual search, yet the effectiveness of cue properties and their failure modes remains underexplored. The current experiments simulate AR overlays on a wide-field-of-view (FOV) display to examine how cue properties affect search performance under perfect or imperfect automation. Across both studies, cues improved search performance relative to an unaided search. Spatial information outperformed identity information, and local cues outperformed global cues. Dual cues outperformed single cues, showing a greater benefit when targets were outside the immediate FOV. Imperfect automation degraded performance and produced an automation bias: Errors were higher when automation was incorrect than with a no-cue baseline (no automation). Miss-prone automation (Experiment 2) yielded larger response time costs than false-alarm-prone automation (Experiment 1), whereas costs to accuracy were comparable. Under miss-prone automation, dual cues slightly amplified the automation bias relative to single cues. These findings have implications for designing automated cueing aids, particularly in large search environments, where spatial precision and automation errors should be considered carefully. (PsycInfo Database Record (c) 2026 APA, all rights reserved).
