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.