Twitter has suspended 125,000 accounts for threatening or promoting terrorist acts, most of them in the cause of Islamic State, the company said on Friday evening.
- Will ISIS' Sophisticated Web Presence Be Its Lasting Legacy?
- How ISIS Fights: Terror, Insurgency and Slick Propaganda
According to CNBC, a previous estimate of ISIS-linked Twitter accounts, prepared by The Brookings Institute, placed their number at about 46,000.
In its statement, Twitter also said it has boosted the teams that review accounts reported by users, and has started to look into accounts similar to those reported. According to Twitter, they have already seen results, "including an increase in account suspensions and this type of activity shifting off of Twitter."
The announcement was especially notable because Twitter has said little about its efforts to combat ISIS and other terror-related content even though it has been criticized for not doing enough to stop those groups from using the service.
ISIS has heavily relied on the 300 million-person site, as well as others, to recruit fighters and propagate violent messages and videos.
Twitter updated its policies last month to explicitly ban "hateful conduct."
The U.S. government has pressured technology companies to cooperate and help them identify terror-related accounts, though Silicon Valley has been wary of engaging with government officials.
In January, a high-profile delegation of top national security officials, including White House Chief of Staff Denis McDonough, Federal Bureau of Investigation Director James Comey and Attorney General Loretta Lynch flew to San Jose to meet tech industry leaders from Twitter, Facebook Inc, Apple Inc , and Google parent Alphabet Inc, but most companies, including Twitter, did not send their chief executive officers.
Still, Twitter said in a blog post that it has cooperated with law enforcement when appropriate, noting that in July 2015 Comey praised Twitter as "very good and thoughtful and hardworking at trying to shut down (terrorism-related) accounts."
"There is no 'magic algorithm' for identifying terrorist content on the internet," Twitter said in its blog post.
It added that it tries to strike a balance between enforcing its rules on prohibited behaviors, the needs of law enforcement and the desire by users to share their views - "including views that some people may disagree with or find offensive."