Hate Speech-Detecting AIs Can Be Fooled: Study

The study found that bad grammar and awkward spelling might make toxic social media comments harder for artificial intelligence (AI) detectors to spot.

from RSS Feeds : RSS Feed - NDTV Gadgets360.com https://ift.tt/2MDzs3W

Comments

Popular posts from this blog

WhatsApp group calls are rolling out to everyone, but don’t ditch Skype just yet

AptX Adaptive is Qualcomm’s latest solution to bad Bluetooth audio

When algorithms go wrong we need more power to fight back, say AI researchers