PulseAugur
中
实时 17:11:22
English(EN) We hid codes in 7 social media posts and asked 5 AI assistants to find them

研究发现人工智能助手难以阅读社交媒体帖子

一项最新研究调查了人工智能助手访问和解读社交媒体帖子内容的能力,揭示了其显著的局限性。研究人员在抖音、小红书、哔哩哔哩、YouTube、TikTok和X等平台上的帖子中嵌入了隐藏的代码,然后要求Gemini、Copilot等人工智能助手查找。虽然大多数人工智能助手因登录墙或验证页面而正确报告无法访问帖子,但Gemini从YouTube上找到了两个代码。开源工具表现稍好,一个名为ReadProof的读取工具实现了最高的成功率,正确识别了九个代码。 AI

影响 强调了人工智能代理需要改进网页浏览能力,以便能够可靠地访问和处理来自各种在线信息源的信息。

排序理由 该项目详细介绍了一项基准研究,评估了人工智能助手访问和解读社交媒体内容的能力,包括发布的一篇论文和开源工具。[lever_c_demoted from research: ic=1 ai=1.0]

在 dev.to — LLM tag 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →

研究发现人工智能助手难以阅读社交媒体帖子

本文如何被排名

Signal score
41 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
该项目详细介绍了一项基准研究,评估了人工智能助手访问和解读社交媒体内容的能力,包括发布的一篇论文和开源工具。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
product, other
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

  1. dev.to — LLM tag TIER_1 English(EN) · Programming with Jack Chew ·

    我们在7条社交媒体帖子中隐藏了代码,并让5款AI助手来寻找它们

    <p>When you paste a Douyin, Xiaohongshu or Bilibili link into an AI assistant, does it actually read the post? You can't tell from the answer. So we planted codes in our own posts and asked.</p> <h2> The test </h2> <p>We published 7 posts on our own accounts on 21 November 2026: …