Of those accounts that were removed between July 1st and December 31st last year, 93 percent were flagged by Twitter's internal tools and algorithms. Around 74 percent were suspended before ever tweeting -- stats that are in line with Twitter's report from the first half of 2017 as well. Government reports of violations are still on the decline. Last September, Twitter reported that government-originating reports accounted for less than one percent of removals. Now, they account for less than 0.2 percent.
In addition to the terrorist content updates, Twitter also shared information on some changes it has made to how it notes when content is being withheld. It's now differentiating when content is being blocked due to a court order or because of local laws. The Transparency Report also notes that there was a 38 percent increase between the first and second half of 2017 regarding copyright violation takedown notices for both Twitter and Periscope. There was an 18 percent decrease in trademark violation notices.
Twitter makes a point to note the efforts of governments to limit online speech. "With the passage of new legislation and ongoing regulatory discussions taking place around the world about the future of public discourse online, we are seeing a potential chilling effect with regards to freedom of expression," Twitter said in a blog post. To address that, Twitter says it uploads request for content withholding to the Lumen database, which collects and analyzes these sorts of requests. "Lumen serves as a critical transparency resource as more freedom of expression comes under fire, by making such requests available for public review," said Twitter.
Lastly, Twitter reiterated its effort to "increase the collective health, openness and civility of public conversation around the world, and to hold ourselves publicly accountable toward progress." In regards to that, Twitter also recently hosted a Periscope livestream, taking questions from the public about its policies, and launched a request for public proposals on how to promote healthy, open and civil conversations on its platform.