Facebook's efforts to stop online extremism on the social network has led the company to ban more than 200 white supremacist groups.
The company also revealed it's been blocking at least some user-posted links to fringe websites 8chan and 4chan, which are known for hosting hate speech. The El Paso shooter posted a document on 8chan prior to last month's mass shooting, for example, prompting networking infrastructure providers Cloudflare and Voxility to terminate services with the site, which remains offline.
Facebook detailed the efforts at a Senate hearing today, where execs from Google and Twitter also testified about efforts to identify and remove violent and extremist content from their platforms. Ahead of the hearing, Facebook provided an update on its attempts to tackle the problem using AI and human-powered content moderation.
In March, the perpetrator behind a separate mass shooting in Christchurch, New Zealand, was able to live stream his attack on Facebook, despite the company's safeguards. Facebook later explained that its automated content moderation simply wasn't smart enough to recognize the disturbing nature of the footage. However, the company now says it has potential solution: Facebook will obtain camera footage from US and UK law enforcements officials on firearms training to teach its AI systems to spot videos depicting mass shootings.
"With this initiative, we aim to improve our detection of real-world, first-person footage of violent events and avoid incorrectly detecting other types of footage such as fictional content from movies or video games," Facebook said.
The same post also mentions Facebook's action to ban more than 200 white supremacist groups, which involved using AI and human-based experts to also remove content praising or supporting the organizations. The takedowns occur as the company has been banning calls for white nationalism and white separatism on the social network. Previously, it permitted both ideologies, classifying them as extreme right-wing political movements. But in March, Facebook began tying the content directly to the promotion of "white supremacy," a subject that was already barred on the platform.
"We'll need to continue to iterate on our tactics because we know bad actors will continue to change theirs, but we think these are important steps in improving our detection abilities," the company added.
In a letter to US Rep. Max Rose (D-NY), Facebook's global policy management head Monika Bickert also said: "Please know that Facebook is blocking links to places on 8chan and 4chan that are dedicated to the distribution of vile content." During today's hearing, Bickert specifically cited 8chan's /Pol board (politically incorrect bulletin board) as one of the blocked domains, though it's unclear when that ban went into effect.
Despite the efforts from Facebook, some experts are still concerned the social network isn't doing enough to stop online extremism. The social network also hasn't identified which white supremacist groups it banned or how many followers they had.
"When Facebook claims in its transparency report that it took action on four million pieces of hate speech from January to March 2019, it is difficult to understand what this means in context," said George Selim, a senior vice president with the Anti-Defamation League, who also testified at today's Senate hearing. He's calling on Facebook to provide more transparency and data on the hate speech takedown efforts that can be verified by third parties.