Mapping public perception of artificial intelligence: Expectations, risk–benefit tradeoffs, and value as determinants for societal acceptance
基于德国1100名全国代表性样本，研究了公众对71个AI场景的期望、风险收益权衡和价值评估，发现风险（β=-0.490）和收益（β=+0.672）解释了96.5%的价值判断差异，为AI开发者和政策制定者提供促进社会接受的见解。
Public opinion on artificial intelligence (AI) plays a pivotal role in shaping trust and AI alignment, ethical adoption, and the development equitable policy frameworks. This study investigates expectations, risk–benefit tradeoffs, and value assessments as determinants of societal acceptance of AI. Using a nationally representative sample (N = 1100) from Germany, we examined mental models of AI and potential biases. Participants evaluated 71 AI-related scenarios across domains such as autonomous driving, medical care, art, politics, warfare, and societal divides, assessing their expected likelihood, perceived risks, benefits, and overall value. We present ranked evaluations alongside visual mappings illustrating the risk–benefit tradeoffs. Our findings suggest that while many scenarios were considered likely, they were often associated with high risks, limited benefits, and low overall value. Regression analyses revealed that 96.5% ( r 2 = 0 . 965 ) of the variance in value judgments was explained by risks ( β = − 0 . 490 ) and, more strongly, benefits ( β = + 0 . 672 ), with no significant relationship to expected likelihood. Demographics and personality traits, including age, gender, and AI readiness, influenced perceptions, highlighting the need for targeted AI literacy initiatives. These findings offer actionable insights for researchers, developers, and policymakers, highlighting the need to communicate tangible benefits and address public concerns to foster responsible and inclusive AI adoption. Future research should explore cross-cultural differences and longitudinal changes in public perception to inform global AI governance.