The ‘original sounds’ of terrorist leaders: A TikTok feature enables terrorist content to flourish
18 July 2024
By: Moustafa Ayad
Ahead of June anniversary of the Islamic State’s (IS) founding of its ‘caliphate,’ ISD undertook analysis to understand how much of the group’s audio content, including from past and current leadership, was easily found on TikTok. Our findings suggest not only are the speeches of the leaders of the group easily found on the platform, but similarly al-Qaeda’s past leadership is easily found through simple searches of their names. Furthermore, even when the original content linked to these ideologues is taken down, TikTok’s ‘original sound’ feature allows the audio to live on the platform through other users’ content. The result was hundreds of videos, all using the ‘original sounds’ of terrorist ideologues proliferating on the platform. Our research builds on previous analysis undertaken on Facebook, which revealed more than 150 pages that posted more than 2,500 pieces of audio content linked to the IS al-Bayan radio station.
It’s a simple read-along TikTok video, coming in just over 2 minutes and 30 seconds. As the narrator drones on, every word spoken in the video appears on screen, highlighted in purple.
The narrator is Abu Hudhayfah al-Ansari, the current spokesman for the Islamic State (IS); the sentence declares that the terrorist group will “renew and repeat [it]s incitement for the lone lions to strive intensely to target the Crusaders and Jews in every place.”
Posted in March 2024 when al-Ansari first delivered the speech, the video has more than 99,700 views and has been liked more than 6,300 times. It is also the primary audio source for seven other videos on TikTok; three are carbon copies of the original video while the remaining four use the audio for different video content. In total, these seven videos added another 43,300 views to the total, and ensured the audio survives even if TikTok removes the original for a clear violation of its Violent and Hateful Organizations and Individuals Community Guidelines.
Al-Ansari’s video is not the only breach of these rules, unfortunately. Between 16-18 June, ISD research found 55 videos on TikTok featuring al-Qaeda and IS ideologues, using a simple randomised search for their names; together, they have received more than 1.1 million views. 60 percent of the videos used ‘original sounds’, which is audio that accompanies a TikTok video and can be used by other creators on the platform in their own videos. 230 other videos on the platform used the audio, gaining an additional 663,793 views.
Al-Ansari’s video was the most viewed in the original dataset. However, a portion of a speech given by Abu Muhammad al-Adnani followed by an IS nasheed — an acapella-like chant — was the most popular for resharing. It was used as the audio track for another 89 videos on the platform. It is important to note that nasheeds, are incredibly popular and most have no affiliation to terror groups.
The primary findings of this research suggest users looking to spread terrorist content through original sounds often use misleading or innocuous titles to circumvent moderation. Using innocuous titles creates wider appeal for the videos, because users are often unaware of the provenance of the speech. Misleading labels also allow videos linked to IS and al-Qaeda to persist even in the event the original video is taken down.
This is not the first time TikTok has faced challenges with identifying and removing IS or al-Qaeda videos. In June 2023, ISD research found that IS content was thriving on TikTok. Issues with the original sounds feature was identified as a clear moderation challenge at the time.
The constant battle against extremist propaganda including audio is not unique to TikTok alone. In January 2024, ISD researchers found 154 Facebook pages which had shared 2,538 videos consisting of speeches from various al-Qaeda and IS ideologues, as well as programmes from the IS al-Bayan radio station. In Facebook’s case, the content was clearly labelled, and the pages were explicit about their origins. By contrast, there remains a consist and persist effort to confound moderators on TikTok about the speaker or organization behind an “original sound.”
Along with al-Ansari and al-Adnani, ISD researchers found portions of speeches by Osama bin Laden, Abu Bakr al-Baghdadi, Abu Omar al-Baghdadi, Abu Musab al-Zarqawi and Turki al-Binali on TikTok. Some videos featured the names or images of the speakers, clearly showing that terrorist content thrives on the platform even when there are clear indicators that it is in violation of TikTok’s guidelines; many others hid behind images of nature or slapdash graphics.
Similarly, users created titles such as “the speech of truth” or naming the ‘original sound’ after the user that originally posted the content, allowing the audio to continue to be spread even in the event of a takedown. One video featuring an IS fighter speaking directly to the camera and addressing the Sunnis of Lebanon generated more than 217,000 views. The video was taken down in 72 hours (about three days) by which time the original sound was already being used by three other videos, which all survived the takedown. One of these was a direct replica of the original, with no alterations or adjustments made to it at the time of its posting.
In the entire dataset of the 55 videos surfaced using a randomised search of ideologue names, just three were taken down as of 18 June, roughly 2 percent. Another review of the accounts undertaken on July 3, indicated 20% of the original videos were taken down, or 12 in total. All the videos which used the original sounds from the videos which were taken down, however, remained on the platform.
Another key finding from the research, is that around 64 of the users (28 percent of the dataset) behind the (230) videos which used the original sounds of terrorist ideologues exhibited no other behaviours to suggest they were supportive of the Islamic State or al-Qaeda. One user used an al-Baghdadi speech over a montage of baby photos, while another used it for filming his children at the playground. In one video, which was set to an al-Zarqawi speech, a mechanic could be seen posing with cars being repaired.
This trend could reflect several possibilities. First, many of the speeches are stripped of their original context. Second, the speeches are often mislabelled, so users believe they are posting content from a non-terrorist source. Third, speeches that were integrated with nasheeds were often the most likely to have more use on the platform than those without.
The research also suggests that the platform’s algorithm provides additional violative content to users already searching for it. This indicates the algorithm does indeed work based on preferences or behaviours of users, and also highlights how TikTok could easily create the conditions for a rabbit hole for other violative content.
The research account for the study was provided recommendations for new terrorist ideologue speeches on the platform not seen by researchers, as well as celebratory videos of accelerationist neo-Nazi terrorists including the Christchurch and Buffalo mass shooters. These videos were served to the account as “videos you may like,” illustrating how the platform’s algorithm could send a potential user further down the rabbit hole of violent extremist propaganda.
This research clearly shows that TikTok’s original sounds feature is a boon for terrorist content on the platform. The feature allows incitement and calls for violence to survive well after they are silenced by their real-world deaths or by take downs.