Twitter has revealed it removed 376,890 accounts promoting terrorism between July and the end of December 2016 – and that the majority were removed by automated technology.
A total of 636,248 accounts have been removed since 1 August 2015, according to the company’s latest transparency report.
Seventy-four percent were identified by proprietary Twitter technology, which automatically scans accounts.
Only 2% were removed after requests by governments.
The technology is an extension of Twitter’s own anti-spam technology, which analyses how an account behaves, rather than what it posts. Accounts promoting terrorism exhibit “distinctive behaviour”.
The report comes as social media companies come under increased scrutiny over how they police their platforms.
Last week, MPs attacked Twitter, Facebook and Google for failing to do enough to tackle online extremism and hate crimes.
Yvette Cooper MP, chair of the influential home affairs committee, said the companies had “a terrible reputation” for failing to act on hate speech and other offensive material.
Google has come under pressure as big brands, including BMW and McDonald’s UK, suspend advertising on its YouTube platform after adverts appeared next to hate speech videos.
Google apologised, with chief business officer Philipp Schindler writing in a blog post: “So starting today, we’re taking a tougher stance on hateful, offensive and derogatory content.”
