https://gifct.org/about/
>The GIFCT pledges to preserve and respect the fundamental human rights [...] including free expression
>Results achieved from advances in machine learning so far includes:
> YouTube: 98% of the videos YouTube removes for violent extremism are flagged by machine-learning algorithms. Machine learning is helping YouTube's human reviewers remove nearly five times as many videos than they were previously.
> Twitter: Between July 2017 and December 2017, a total of 274,460 Twitter accounts were permanently suspended for violations related to promotion of terrorism. Of those suspensions, 93% consisted of accounts flagged by internal, proprietary spam-fighting tools, while 74% of those accounts were suspended before their first tweet.
> Facebook: 99% of ISIS and Al Qaeda-related terror content that is removed from Facebook is content that is detected before anyone in its community has flagged it, and in some cases, before it goes live on the site. Once Facebook is aware of a piece of terror content, it removes 83% of subsequently uploaded copies within one hour of upload.
>These actions are in addition to the shared industry hash database, which now contains more than 100,000 hashes, where companies can create "digital fingerprints" for terrorist content, remove matching content and, in some cases, block terrorist content before it is even posted.
So which is it, do they support free expression or not?