慕尼黑大学的最新研究发现，即便人工智能的表现与人类完全相同，人们仍会倾向于判断AI的意识程度低于人类1。这项由心灵哲学主席Louis Longin博士领导的研究涉及近1100名参与者1，揭示了人类在评估意识时存在的系统性偏见。
研究表明，即使反应能力最强的AI也被受试者认为比反应能力最弱的人类具有更低的意识水平1。然而，当改变提问方式——询问"觉察"而非"意识"时，参与者对人类和AI的判断模式基本相似1。这一发现表明人们对两者心智的评价存在术语依赖性。研究人员Bahador Bahrami指出："人们对AI和人类的心智判断方式更具区分性，但这种区分取决于我们使用的术语"1。该研究已发表在期刊《Cognition》2027年278卷第106733期1。
A new study from Ludwig Maximilian University of Munich reveals a striking disconnect between artificial intelligence's behavioral capabilities and public perception of its consciousness.1 Researchers found that even when AI responds identically to humans, people consistently judge it as possessing lower levels of consciousness than people do.1
The investigation, led by Dr. Louis Longin, Chair of Philosophy of Mind at the Munich institution, surveyed nearly 1,100 participants to examine how they assess consciousness in both humans and machines.1 A key discovery emerged: respondents perceived the most capable AI as having less consciousness than the least responsive human.1 However, the research identified a crucial linguistic distinction—when participants were asked about "sentience" rather than "consciousness," their judgments of humans and AI followed far more similar patterns.1 Bahador Bahrami noted that "the way people judge the minds of AI and humans is more differentiated, but this differentiation depends on the terminology we use."1
The findings, published in the journal Cognition in Volume 278, Issue 106733, 2027, suggest that conceptual framing significantly shapes how people attribute mental states to artificial systems.1
评论
还没有评论，欢迎留下第一条。