Embed Notice

HTML Code

<blockquote style="position: relative; padding-left: 55px;"><section><a href="https://toad.social/users/KimPerales/statuses/114536276007365947">Kim Perales (kimperales@toad.social)'s status on Tuesday, 20-May-2025 05:23:53 JST</a><a href="https://toad.social/@KimPerales" title="kimperales@toad.social"><img src="https://gnusocial.jp/avatar/119120-48-20250313211700.webp" width="48" height="48" alt="Kim Perales" style="position: absolute; left: 0; top: 0;">Kim Perales</a></section><article><p>"New study reveals that when used to summarize scientific research,🚨generative AI is nearly five times LESS accurate than humans.</p><p>Many haven't realized, but Gen AI's accuracy problem is worse than initially thought."<br>-L Jarovsky</p><p>Many are realizing that generative AI is overhyped &amp; it underdelivers.<br><a href="https://toad.social/tags/Technology" rel="tag">#Technology</a> <a href="https://toad.social/tags/AI" rel="tag">#AI</a> <a href="https://toad.social/tags/Fraud" rel="tag">#Fraud</a> <a href="https://toad.social/tags/USPol" rel="tag">#USPol</a> <a href="https://royalsocietypublishing.org/doi/epdf/10.1098/rsos.241776" rel="nofollow noreferrer">https://royalsocietypublishing.org/doi/epdf/10.1098/rsos.241776</a></p></article><footer><a rel="bookmark" href="https://gnusocial.jp/conversation/5077150#notice-9955735">In conversation</a><time datetime="2025-05-20T05:23:53+09:00" title="Tuesday, 20-May-2025 05:23:53 JST">about 6 days ago</time> <span>from <span><a href="https://toad.social/@KimPerales/114536276007365947" rel="external" title="Sent from toad.social via ActivityPub">toad.social</a></span></span><a href="https://toad.social/@KimPerales/114536276007365947">permalink</a><h4>Attachments</h4><ol><li><label><a rel="external" href="https://gnusocial.jp/attachment/4651500">The image is a research article titled "Generalization bias in large language model summarization of scientific research" by Uwe Peters and Benjamin Chin-Yee. It includes the authors' affiliations, publication details, and a summary discussing the potential and limitations of…</a></label><br><a href="https://cdn.toad.social/media_attachments/files/114/536/272/179/705/478/original/3e7e7802c46cc355.png" rel="external">https://cdn.toad.social/media_attachments/files/114/536/272/179/705/478/original/3e7e7802c46cc355.png</a></li><li><label><a rel="external" href="https://gnusocial.jp/attachment/4651501">Untitled attachment</a></label><br></li></ol></footer></blockquote>

Corresponding Notice

Embed this notice
Kim Perales (kimperales@toad.social)'s status on Tuesday, 20-May-2025 05:23:53 JST Kim Perales
"New study reveals that when used to summarize scientific research,🚨generative AI is nearly five times LESS accurate than humans.
Many haven't realized, but Gen AI's accuracy problem is worse than initially thought."
-L Jarovsky
Many are realizing that generative AI is overhyped & it underdelivers.
#Technology #AI #Fraud #USPol https://royalsocietypublishing.org/doi/epdf/10.1098/rsos.241776
In conversationabout 6 days ago from toad.socialpermalink
Attachments
1. The image is a research article titled "Generalization bias in large language model summarization of scientific research" by Uwe Peters and Benjamin Chin-Yee. It includes the authors' affiliations, publication details, and a summary discussing the potential and limitations of…
  https://cdn.toad.social/media_attachments/files/114/536/272/179/705/478/original/3e7e7802c46cc355.png
2. Untitled attachment

Public

Embed Notice

HTML Code

Corresponding Notice