acm-header
Sign In

Communications of the ACM

ACM Opinion

Facebook Uses Deceptive Math to Hide Its Hate Speech Problem


View as: Print Mobile App Share: Send by email Share on reddit Share on StumbleUpon Share on Hacker News Share on Tweeter Share on Facebook
Young woman on smartphone appears to be upset.

Technically, both numbers are correctthey just measure different things. The measure that really matters is the one Facebook has been hiding.

Getty Images

Amid the thousands of Facebook documents leaked by whistleblower Frances Haugen, is a letter from March of this year stating the company removes "as little as" 3% to 5% of the hate speech on the site, though it consistently claims removal rates of more than 90%.

Facebook's utter failure to limit hate speech and the simple deceptive trick it's consistently relied on to hide this failure is shocking. It exposes just how much Facebook relies on artificial intelligence (AI) for content moderation, just how ineffective that AI is, and the necessity to force Facebook to come clean.

From Wired
View Full Article


 

No entries found