PulseAugur
EN
LIVE 09:54:36

Bluesky moderation system shows high precision but low recall in harm detection

By PulseAugur Editorial · [1 sources] ·

A new study published on arXiv analyzes the Bluesky Moderation Service (BMS), examining 10.6 million moderation labels from 2025. The research reveals that BMS operates as a human-AI collaborative system, with automated labeling for sexual and graphic content occurring rapidly, while more complex issues require human oversight and take longer. The study found BMS has high precision (0.837) but low recall (0.222), meaning it accurately flags content it identifies but misses a significant amount of harmful material. The analysis also identified detected harms including hostility towards protected groups and the dissemination of explicit content. AI

IMPACT Provides insights into the operational effectiveness of AI-assisted content moderation systems on social platforms.

RANK_REASON Research paper published on arXiv detailing an audit of a social media platform's moderation service. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

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

Bluesky moderation system shows high precision but low recall in harm detection

How we ranked this

Signal score
12 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
Research paper published on arXiv detailing an audit of a social media platform's moderation service. [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
paper, safety
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. arXiv cs.AI TIER_1 English(EN) · Pushpdeep Singh, Sayeh Jarollahi, Ayan Majumdar, Vabuk Pahari, Abhijnan Chakraborty, Krishna P. Gummadi, Ingmar Weber, Abhisek Dash ·

    Characterizing Bluesky Content Moderation Service: From Automation of Service to Landscape of Harms

    arXiv:2609.11373v1 Announce Type: cross Abstract: Empirical research on content moderation is fundamentally constrained by the opaque deployment of moderation systems on major social media platforms. To this end, the recent emergence of decentralized platforms with transparent, p…

RELATED ENTITIES

RELATED TOPICS