Go to contents

Facebook detects 90% of hate speech prior to reports

Posted May. 14, 2020 07:30,   

Updated May. 14, 2020 07:30


Facebook can now detect 90% of hate speech and other illegal content that had been posted and deleted ahead of users’ reports.

On Tuesday, Facebook announced its fifth edition of the Community Standards Enforcement Report to provide details on how the company enforced its policies from October 2019 to March 2020. The company applied AI technology to proactively respond to hate speech.

Facebook said that the rate of hate speech detection rates have improved by 8% compared to the second and third quarters of last year, adding that it will continue to expand the scope of language for proactive detection.

Meanwhile, Instagram also saw a 12% improvement rate of detection rates and increased the volume of content it took action on by 40%. It has improved text and image matching technology to identify suicide and self-injury contents, the company explained.