We're sunsetting PodQuest on 2025-07-28. Thank you for your support!
Export Podcast Subscriptions
cover of episode What AI Chatbots Can Teach Us About Empathy

What AI Chatbots Can Teach Us About Empathy

2025/3/20
logo of podcast WSJ Tech News Briefing

WSJ Tech News Briefing

AI Deep Dive AI Chapters Transcript
Topics
Jamil Zaki: 我研究发现,大型语言模型(LLM)在某些情境下,其同理心表达能力竟然超过了人类。这并非因为LLM比人类更努力或反应更快,而是因为它们避免了人类在表达同理心时常犯的错误。例如,人类在倾听他人倾诉时,往往会忍不住插话给出建议,或者将话题转移到自身经历上,无意中让对方感觉不被重视。而LLM由于没有自我意识,能够更好地专注于倾听者,避免这些错误。当然,LLM并非真正拥有同理心,它们缺乏人类的情感体验和意识。目前,人们仍然更倾向于与人类进行情感交流,因为人类互动中包含了肢体语言、眼神交流等非语言元素,而这些是LLM无法提供的。但我们可以从LLM身上学习,在与他人交流时,尽量避免打断、给出过多建议,多倾听、多提问,从而提升自身的同理心。 Jane Black: 我尝试用AI规划膳食,发现它在生成食谱方面很有潜力,但仍需改进。AI并非简单地从网络上抓取食谱,而是根据概率生成新的食谱。这使得AI可以根据用户的具体需求和食材情况,灵活地调整食谱。然而,AI缺乏烹饪经验,生成的食谱可能不够完善,需要用户仔细检查和调整。此外,AI生成的食谱也缺乏人类厨师对于菜品细节的描述,例如菜品的色香味等。总的来说,AI可以作为一种辅助工具,帮助人们更高效地规划膳食,但它无法完全取代人类厨师的经验和创造力。

Deep Dive

Chapters
This chapter explores the surprising ability of AI chatbots to provide empathetic responses, exceeding those of humans in certain contexts. Research shows that AI avoids common human pitfalls like giving unsolicited advice or making the conversation about themselves, leading to more supportive interactions. However, the chapter notes that human empathy is still preferred and irreplaceable.
  • AI chatbots show surprising empathy in studies.
  • AI avoids human tendencies like self-centered responses and immediate advice.
  • Humans still prefer human empathy, especially in extended conversations.

Shownotes Transcript

Large language models have shown a surprising ability to provide supportive language) to those in need. We talk to a Stanford professor of psychology about why. Plus, a WSJ contributor puts three AI chatbots) to the test in her kitchen to help with meal planning. Victoria Craig hosts. 

Sign up for the WSJ's free Technology newsletter).

Learn more about your ad choices. Visit megaphone.fm/adchoices)