Twitter expands efforts in AI-assisted warfare to false COVID information


From July 1 to December 6, Twitter deleted a complete of three.8 million tweets for violating its guidelines (not simply these associated to COVID-19), and decided that 77 % of those tweets have been considered lower than 100 instances previous to elimination. Nonetheless 6 % (or 228,000 tweets) have been considered greater than a thousand instances earlier than they have been eliminated.


“Our purpose is to enhance these numbers over time, taking enforcement motion on violent content material earlier than it’s even seen,” he stated on Twitter.

Elsewhere within the Twitter transparency report, the corporate stated it had taken motion in opposition to almost a million accounts for abuse and harassment in July-December 2020 (up 142 % from the earlier half). The class of violent tweets that noticed the most important enhance was “nudity with out consensus,” with Twitter reporting actions in opposition to 27,087 accounts for a 194 % enhance.

There was a 35 % lower within the variety of accounts completely suspended for terrorist or violent extremist messages, and nearly all have been proactively recognized with the assistance of AI.


Please enter your comment!
Please enter your name here