PulseAugur
EN
LIVE 17:03:16

AI assistants struggle to read social media posts, study finds

A recent study investigated the ability of AI assistants to access and interpret content from social media posts, revealing significant limitations. Researchers embedded hidden codes within posts on platforms like Douyin, Xiaohongshu, Bilibili, YouTube, TikTok, and X, then tasked AI assistants such as Gemini, Copilot, and others with finding them. While most AIs correctly reported inability to access posts due to login walls or verification pages, Gemini managed to find two codes from YouTube. Open-source tools showed slightly better performance, with a reading harness called ReadProof achieving the highest success rate by correctly identifying nine codes. AI

IMPACT Highlights the need for improved web-browsing capabilities in AI agents to reliably access and process information from diverse online sources.

RANK_REASON The item details a benchmark study evaluating AI assistant capabilities in accessing and interpreting social media content, including the release of a paper and open-source harness. [lever_c_demoted from research: ic=1 ai=1.0]

Read on dev.to — LLM tag →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

AI assistants struggle to read social media posts, study finds

How we ranked this

Signal score
43 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
The item details a benchmark study evaluating AI assistant capabilities in accessing and interpreting social media content, including the release of a paper and open-source harness. [lever_c_demote…
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.

Full methodology in our editorial standards.

COVERAGE [1]

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

    We hid codes in 7 social media posts and asked 5 AI assistants to find them

    <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: …