Bluesky Transparency Report: Content Labeling Surges 200%

Bitcoin reaches new all-time high of over $118,000 within 24 hours

Bitcoin reached a new all-time high of $118,900 on Friday, surpassing its previous record of $113,822 set on Thursday. As Read more

Conveyor Revolutionizes Vendor Security Reviews and RFPs with AI

Selling software to companies can be a daunting task, especially when it comes to meeting security requirements. Chas Ballew, founder Read more

Ready-made Stem Cell Therapies in Development for Pets

Earlier this week, San Diego startup Gallant announced $18 million in funding to bring the first FDA-approved ready-to-use stem cell Read more

Elon Musk’s Dodgy Election Claims Have Gone Viral with 2 Billion Views on X

The world’s richest man buys out one of the most popular social media platforms and uses it as a propaganda Read more

 

 

Bluesky released its inaugural transparency report for 2025, detailing a massive scale-up in moderation enforcement and legal compliance as user reports and platform activity reached record levels. The document outlines how the decentralized social network is managing growth while combatting spam, harassment, and influence operations.

Analysis of User Reports and Moderation Categories

The report highlights a complex landscape of user-generated reports. A broad “other” category accounted for 22.14% of total reports, covering issues that did not specifically fall under violence, child safety, or self-harm. Within the “misleading” category, which saw 4.36 million reports, spam was the primary driver with 2.49 million instances.

Harassment and Antisocial Behavior

Harassment remains a significant focus for the platform, totaling 1.99 million reports. Hate speech represented the largest specific share within this group at 55,400 reports, followed by targeted harassment (42,520), trolling (29,500), and doxxing (3,170). Bluesky noted that a majority of harassment reports involved “antisocial behavior”—rude remarks that, while unpleasant, do not always violate specific policies like hate speech.

See also  GenAI Overload: Focus on Specific Goals

Sexual Content and Metadata Accuracy

Regarding sexual content, the platform received 1.52 million reports, though most concerned mislabeling rather than prohibited material. Bluesky reported that adult content often lacked the necessary metadata tags required for users to filter their own experience. More severe violations included nonconsensual intimate imagery (7,520), abuse content (6,120), and deepfakes (over 2,000).

Automated Systems and Violence Prevention

Reports concerning violence totaled 24,670, categorized into threats or incitement (10,170), glorification of violence (6,630), and extremist content (3,230). Beyond manual reporting, Bluesky’s automated systems proactively flagged 2.54 million potential violations.

Efficiency gains were noted following the implementation of a system that hides toxic replies behind an extra click. This change resulted in a 79% drop in daily reports of antisocial behavior. Overall, reports per 1,000 monthly active users declined by 50.9% between January and December.

Takedowns, Legal Requests, and State Actors

The platform took decisive action against foreign interference, removing 3,619 accounts linked to suspected influence operations, primarily originating from Russia. This aligns with Bluesky’s stated goal of becoming more aggressive regarding enforcement.

Growth in Enforcement Actions

In 2025, Bluesky removed 2.44 million items, including content and entire accounts. This represents a massive increase from the previous year, where only 66,308 accounts were removed. Automated tools were responsible for 35,842 of those prior takedowns, while manual moderators removed 6,334 records.

Suspensions vs. Labeling Strategy

Bluesky issued 3,192 temporary suspensions and 14,659 permanent removals for ban evasion in 2025. Permanent bans were largely concentrated on inauthentic behavior, spam networks, and impersonation. Despite these figures, the data suggests a strategic preference for labeling over account termination. Content labels rose 200% year-over-year to 16.49 million, while account takedowns grew by 104% (from 1.02 million to 2.08 million). Most labels were applied to nudity and suggestive adult content.

See also  OpenAI cautions against SPVs and other 'unauthorized' investments

Bluesky 2026 Roadmap: Real-Time Features and Feed Upgrades

Wayback Machine Plugin Fixes WordPress Broken Links