{"id":11258,"date":"2024-05-26T09:23:49","date_gmt":"2024-05-26T01:23:49","guid":{"rendered":"https:\/\/www.1ai.net\/?p=11258"},"modified":"2024-05-26T09:23:49","modified_gmt":"2024-05-26T01:23:49","slug":"%e3%80%8a%e8%87%aa%e7%84%b6%e3%80%8b%e6%9d%82%e5%bf%97%e7%a0%94%e7%a9%b6%ef%bc%9aai-%e8%bf%bd%e8%b8%aa%e4%bb%96%e4%ba%ba%e5%bf%83%e7%90%86%e7%8a%b6%e6%80%81%e8%83%bd%e5%8a%9b%e4%b8%8e%e4%ba%ba","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/11258.html","title":{"rendered":"Nature magazine study: AI&#039;s ability to track other people&#039;s mental states is comparable to or even better than humans"},"content":{"rendered":"<p data-vmark=\"6d01\">In the latest issue of this month,<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e8%87%aa%e7%84%b6\" title=\"_Other Organiser\" target=\"_blank\" >nature<\/a>-Human Behavior&quot; journal, published an article about AI<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e7%a0%94%e7%a9%b6%e8%ae%ba%e6%96%87\" title=\"[Sees articles with labels]\" target=\"_blank\" >Research Papers<\/a>, which mentioned that in the task of testing the ability to track the mental state of others, two types of AI <a href=\"https:\/\/www.1ai.net\/en\/tag\/%e5%a4%a7%e8%af%ad%e8%a8%80%e6%a8%a1%e5%9e%8b\" title=\"[View articles tagged with [large language model]]\" target=\"_blank\" >Large Language Model<\/a>In certain situations, they can perform similarly to or even better than humans.<\/p>\n<p data-vmark=\"f5fc\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-11259\" title=\"922cfa06-a690-44f4-a061-479fa1e49954\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2024\/05\/922cfa06-a690-44f4-a061-479fa1e49954.jpg\" alt=\"922cfa06-a690-44f4-a061-479fa1e49954\" width=\"640\" height=\"427\" \/><\/p>\n<p>Image source: Pixabay<\/p>\n<p data-vmark=\"6b96\">As the key to human communication and empathy, mental state ability (also known as theory of mind) is very important for human social interaction. The first author of the paper, James WA Strachan of the University Medical Center Hamburg-Eppendorf in Germany, together with colleagues and collaborators, selected tasks that can test different aspects of theory of mind.<span class=\"accentTextColor\">Including detecting misconceptions, understanding indirect speech, and identifying faux pas<\/span>.<\/p>\n<p data-vmark=\"dca2\">The team selected the GPT and LLaMA2 models for experiments and compared them with 1,907 people.<\/p>\n<p data-vmark=\"53c7\">The results show that the GPT model is good at identifying indirect requests, false ideas, and misleading<span class=\"accentTextColor\">Able to reach and sometimes even exceed the average human level<\/span>, while the performance of LLaMA2<span class=\"accentTextColor\">Below human level<\/span>; LLaMA2 is better than humans at identifying faux pas, but GPT performs poorly.<\/p>\n<p data-vmark=\"586a\">According to China News Service, the author said that LLaMA2&#039;s success was shown to be due to a low &quot;level of bias&quot; in responses, not a true sensitivity to faux pas, and that GPT&#039;s &quot;poor performance&quot; was due to an &quot;ultra-conservative&quot; attitude toward sticking to conclusions, rather than errors in reasoning.<\/p>","protected":false},"excerpt":{"rendered":"<p>In the latest issue of the journal Nature Human Behavior later this month, a research paper on AI was published, which mentioned that two types of AI macrolanguage models have the ability to perform similarly to or even surpass humans in specific situations in a task that tests the ability to track others' mental states. Pixabay As a key to human communication and empathy, mental state competence (also known as theory of mind) is important for human social interactions. The paper's first author, James W. A. Strachan of the University Medical Center Hamburg-Eppendorf in Germany, along with colleagues and collaborators, chose tasks that would test different aspects of the theory of mind, including detecting false ideas, understanding indirect speech, and recognizing rudeness. The team<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[146],"tags":[706,2760,2761],"collection":[],"class_list":["post-11258","post","type-post","status-publish","format-standard","hentry","category-news","tag-706","tag-2760","tag-2761"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/11258","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=11258"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/11258\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=11258"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=11258"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=11258"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=11258"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}