The problems with AI based censorship

The problems with AI based censorship

A live Youtube discussion video about chess – the game – was censored and taken off line by Youtube’s hate speech detection bots.

It appears that Youtube’s artificial intelligence is closer to artificial stupidity. It detected discussion of “white” and “black” pieces on the board, with phrases including “attack” and “defenses” and concluded the video was hate speech or worse, and censored it.

A research project suggest 80% of AI-driven censorship are false positives – because the AI models lack context.

Source: Why a YouTube Chat About Chess Got Flagged for Hate Speech | WIRED

Comments are closed.