我应该陈述还是展示?使AI与人类偏好对齐
Should I State or Should I Show? Aligning AI with Human Preferences
AI总结:
通过在线实验比较AI代理从陈述偏好(文字提示)和显示偏好(选择数据)中学习人类偏好的效果,发现显示偏好数据预测更准确,但用户常选择信息量较少的方式,且AI在冲突时更倾向于遵循提示。
摘要:
随着AI代理变得更加自主,将其目标与人类偏好正确对齐变得越来越重要。我们研究了AI代理在风险选择中通过陈述偏好与显示偏好学习人类委托人偏好的有效性。我们进行了一项在线实验,受试者通过书面指令(“提示”)陈述其偏好,并通过一系列二元彩票问题中的选择(“数据”)显示其偏好。我们发现,平均而言,给定显示偏好数据的AI代理预测受试者选择的准确性高于给定陈述偏好提示的AI代理。进一步分析表明,这种差距源于受试者难以将自己的偏好转化为书面指令。当被选择向AI代理提供哪种信息源时,大部分受试者未能选择信息量更大的那个。此外,当两个来源的预测冲突时,我们发现AI代理更频繁地与提示对齐,尽管其准确性较低。总体而言,这些结果突出了显示偏好方法作为向AI代理传达人类偏好的强大机制,但其成功取决于谨慎的实施。
英文摘要:
The proliferation of AI agents introduces a new principal-agent problem which stems from human principals' difficulty in articulating preferences. We report results from an experiment focused on mitigating this problem via revealed preferences from choice data. Compared to stated preference from human-written prompts, individuals communicate their preferences more effectively through choices, with as few as two sufficing to match prompts' predictive value. This gap largely reflects subjects' difficulty in translating preferences into prompts and is largest among those exhibiting more Allais-type behavioral patterns. Subjects also misperceive the approaches' relative performance, often choosing the less accurate one.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。