Increased Criticism Over Misuse Of User Data Against Backdrop Of More Apologies & Empty Rhetoric

A U.K. government office announced this week that it intended to fine Facebook more than $600,000 for the Cambridge Analytica data breach, which violated Britain’s data-privacy laws. It gets even worse. Facebook admitted on Tuesday to giving a Russian Internet company with ties to the Kremlin access to user data without their consent.  Facebook claims a policy change stopped the data collection, but as the Counter Extremism Project has documented, time and again these reactionary measures only occur after the damage has already been done.

Facebook’s response to the recent criticism is familiar terrain for the technology giant – touting of its supposed successes and a promise to do better.  Despite a mountain of evidence that Facebook has not even made a dent in the problem of online extremism, the company continues to state its Artificial Intelligence (AI) has removed 99 percent of terrorist-related content from its platforms.  The difference this time, though, is that Facebook now claims the AI can identify terrorist content with 99 percent accuracy – a statement that seems to have shifted the goal post.

It’s obvious Facebook’s reliance on AI isn’t working, nor are past tactics.  As the company faces more backlash for its misuse of user data, a radicalized Virginia man was sentenced last week to 20 years in prison for his support of ISIS.  The Federal Bureau of Investigation (FBI) found that he had been primarily radicalized on social media – notably Facebook – and had researched how to make bombs and purchase weapons online. 

Such content of terrorist propaganda and inciting violence can still be found on the platform today – many with thousands of views.  An ISIS video called “The Point of Death” shows combat and executions of prisoners.  Another ISIS video that originally appeared in February depicts a wheelchair-bound man detonating a bomb.  A video from Ansar Alkhelafa Europe celebrates recent terror attacks and depicts beheadings.  British neo-Nazi group Scottish Dawn released a propaganda video despite being banned by the U.K. government.  These examples must be in that one percent Facebook’s AI can’t catch.

To read more about this issue and see the aforementioned examples, as well as other examples of extremist content, please see the background below.

EXTREMIST FACEBOOK CONTENT​

1. ISIS Video With Over 1,000 Views

• Located on Facebook: July 11, 2018
• ​Time on Facebook: One week
• Views: 1,000, 25 likes/reacts and five share
• URL: Link
• Description: ISIS video “The Point of Death,” originally released on July 4, 2018.  The video, from the group’s self-proclaimed state in Libya, shows combat against unidentified forces, vehicle-borne suicide bombings and several executions of prisoners.

2. Video Segment From ISIS Video With Over 3,000 Views

• Located on Facebook: July 11, 2018
• ​Time on Facebook: One week
• Views: 3,400, 233 likes/reacts and 83 shares
• URL: Link
• Description: A clip from the ISIS video “Inside the Caliphate 7,” originally released in February 2018.  The video segment features a wheelchair bound suicide bomber who is filmed saying goodbye to his daughter and nephew before detonating an explosives-laden vehicle.

3. Fan Made ISIS Video With Over 1,200 Views

• Located on Facebook: July 11, 2018
• Time on Facebook: Approximately one week
• Views: 1,200, 97 likes/reacts and 35 share
• URL: Link
• Description: A fan made ISIS video accompanied by a nasheed, containing various pro-ISIS images and propaganda photos.

4. Pro-ISIS Video Celebrating Attacks In Europe & Including Beheadings With Over 2,000 Views

• Located on Facebook: July 11, 2018
• Time on Facebook: Four days
• Views: 2,200 views, 167 likes/reacts and 82 shares
• URL: Link
• Description: A video made by the pro-ISIS Ansar Alkhelafa Europe forum group titled, “The Battle Has Its Ups and Downs.”  The video contains footage from official ISIS videos and celebrates attacks in Paris, Brussels, Berlin, London, Barcelona, Manchester, Nice and Stockholm.  The video includes news footage from several of those attacks and concludes with the beheadings of three men.  The concluding execution footage was taken from an unidentified previously released ISIS video.

5. Propaganda Video From Banned British Neo-Nazi Group Scottish Dawn

• Located on Facebook: July 11, 2018
• Time on Facebook: Approximately one year
• Views: 3,000, 58 likes/reacts and 59 shares
• URL: Link
• Description: A propaganda video from the banned British neo-Nazi group Scottish Dawn.  The video alleges a conspiracy by European governments to allow refugees to enter and urges people to “resist the invasion” and “support the resistance.”  Scottish Dawn was outlawed by the British government in September 2017, authorities alleged that the group was another name for National Action, a neo-Nazi group banned by the British government in December 2016.

BACKGROUND

Facebook Faces A Massive Fine From The U.K. Over The Cambridge Analytica scandal. “A U.K. government office that investigated the Cambridge Analytica scandal announced its intention to fine Facebook 500,000 pounds ($663,000) for failing to safeguard that user information.  The amount is the maximum that the agency, the Information Commissioner’s Office, can levy for violation of Britain’s data-privacy laws.” (“Facebook Faces U.K. Fine Over Its Privacy Scandal,” The Associated Press, 7/11/18)

Facebook Admitted To Giving A Russian Internet Company, With Ties To The Kremlin, Access To User Data Without Their Consent. “A Russian internet company with links to the Kremlin was among the firms to which Facebook gave an extension which allowed them to collect data on unknowing users of the social network after a policy change supposedly stopped such collection.” (Donie O. Sullivan, Drew Griffin & Curt Devine, “Russian Company Had Access To Facebook User Data Through Apps,” CNN, 7/11/18)

Facebook Now States That “AI Can Identify Terrorism-Linked Content With 99 Percent Accuracy.” “[Facebook Vice President of Ads & Business Platforms Mark] Rabkin says that AI can identify terrorism-linked content with 99 percent accuracy, but hate speech is different.  Facebook’s AI flagged hate speech only 38 percent of the time.  That’s because the machines have difficulty distinguishing between, say, someone describing a hate incident and someone preaching hate.  ‘AI is getting smarter.  It is learning,’ Rabkin says, adding that a year ago AI could identify only about 10 percent of hate speech.” (Garett Sloane, “Facebook Pursues AI In Bid To ID Harmful Content,” Ad Age, 7/9/18)

In January, Officials Arrested Virginia Man For His Ties To ISIS And Stated His Facebook Use Played A Part In His Radicalization. “The U.S. Attorney’s Office detailed in a court affidavit that [Sean] Duncan’s Twitter and Facebook accounts along with an encrypted messaging app were a key part of their investigation that showed his support for ISIS.  They said Duncan’s Twitter handle translated from Arabic means ‘the Islamic State.’” (Tom Fitzgerald, “Sterling Man Accused Of Obstructing ISIS Terror Investigation After Home Raided By FBI,” Fox 5 DC, 1/2/18)

Man Was Sentenced Last Week To 20 Years In Prison. “When the FBI knocked on his front door in December, Sean Duncan says, he ran out the back carrying a thumb drive full of child pornography.  He broke it in half and doused it in cleaning solution.  But in a sign of how strange his criminal case has been, prosecutors say he might be lying in an effort to conceal ties to terrorism or information on the death of his son.  U.S. District Judge Leonie M. Brinkema on Friday sentenced Duncan to 20 years in prison and a lifetime of supervision, the maximum allowed under his plea agreement to charges of obstruction of justice and receipt of child pornography.” (Rachel Weiner, “‘I Was A Poser,’ Says Young Man Caught With Islamic State Propaganda And Child Porn,” The Washington Post, 7/6/18)

###