Meta says it has cut US content removal mistakes by half since its January policy changes, without broadly exposing users to more offensive content than before
Meta says loosening its enforcement policies earlier this year led to fewer erroneous takedowns on Facebook and Instagram …
Wired Paresh Dave
Related Coverage
- Meta says it disrupted influence operations linked to China, Iran, Romania The Record · Daryna Antoniuk
- Integrity Reports, First Quarter 2025 Meta
- Meta publishes its Q1 integrity reports following its hate speech policy changes; Facebook saw an uptick in violent content, bullying, and harassment Engadget · Karissa Bell
- Meta Says it Halved Takedown Errors After Moderation-Policy Shift WinBuzzer · Markus Kasanmascheff
- Meta Disrupts Influence Ops Targeting Romania, Azerbaijan, and Taiwan with Fake Personas The Hacker News
- Meta says online harassment is up and false flags are down following a change in content moderation policies Business Insider · Lakshmi Varanasi
- Meta Claims a 50% Reduction in ‘Enforcement Errors’ Since End of 2024 Breitbart · Sean Moran
- Facebook's content moderation ‘happens too late,’ says new Northeastern research Northeastern Global News · Cyrus Moulton
- It's interesting how they frame this. Thier violent content increased from 0.06% to 0.09% (an increase of 0.03%, or a 50% increase from 0.06%). Meanwhile their moderation mistakes decreased by that same 50%, but it's presented as 50% instead of the other way. @thebrianpenny · Brian Penny
- I have mad respect for you and you're 10x smarter than me, but I really struggle with this framing. A reduction in enforcement mistakes is tautological. You're turning up the dial on automated enforcement thresholds and removing less stuff, so yeah, you have less false positives. … @drewpusateri · Drew Pusateri
- something else that stood out to me was just how much less content Meta is removing after all since it moved away from proactive enforcement. Spam removals on Facebook for example declined by nearly half between Q4 2024 and Q1 2025 @karissabe · Karissa
- Meta has published the first of its quarterly integrity reports following its big content moderation pivot in January. Facebook saw increases in violent content and bulling and harassment but the company also says its cut moderation mistakes in half overall + lots of other really interesting details ⬇ … @karissabe · Karissa
- Facebook sees rise in violent content and harassment after policy changes — Violent content on Facebook jumped from 0.07% to 0.09% and harassment from 0.07% to 0.08%. — Meta took down 3.4M hate speech posts—its lowest since 2020 and nearly half of last year's total. … @mattnavarra · Matt Navarra
Discussion
-
@claesdevreese
Claes de Vreese
on bluesky
New transparency report from Meta following content moderation changes — This begs for actual data transparency: “according to its own sampling, Meta estimates that users were exposed to about one to two pieces of hateful content on average for every 10,000 posts viewed” — ww…
-
@peard33
Paresh Dave
on bluesky
Meta removed nearly one-third less content on Facebook and Instagram globally for violating its rules from January to March of this year than it did in the previous quarter, or about 1.6 billion items compared to just under 2.4 billion, www.wired.com/story/meta-c...
-
@thebrianpenny@threads.com
Brian Penny
on threads
@thebrianpenny@threads.com boosted the following from https://www.threads.com/@karissabe/ post/DKP2YipSdCL — Meta has published the first of its quarterly integrity reports following its big content moderation pivot in January. …