Meta accelerates AI handling of content moderation by year end
Technology
Meta (the company behind Facebook and Instagram) is fast-tracking its plan to let AI handle most content moderation.
By the end of this year, up to 90% of moderation in some areas could be done by large language models instead of people.
The goal? Make things faster, more accurate, and a bit cheaper.
Employees question AI accuracy, contractors cut
Meta says its new AI system already makes fewer mistakes than humans and spots more rule-breaking posts.
But not everyone's convinced: employees worry that AI still misses sarcasm and evolving internet slang, causing accidental post removals or shadow banning.
Plus, with automation ramping up, Meta is ending contracts with outside moderation teams, so the whole way content gets policed is getting a major shake-up.