The video, showing an episode of “Family Guy,” one of the world's most famous cartoon comedy shows, shows the main character, Peter Griffin, driving a van loaded with a bomb across a bridge at gunpoint.
The clip is from the episode “Turban Cowboy,” which aired in 2013. However, the audio accompanying the video has been altered, with Peter Griffin, normally voiced by show creator Seth MacFarlane, singing the unusual lyrics.
“Our weapons are heavy and our soldiers are many, but Allah's soldiers are well prepared,” he sings in a distinctive Rhode Island accent meant to inspire supporters of the Islamic State militant group.
The animation is clearly not McFarlane's latest satirical piece, but just one example of how extremist groups are using advanced computing and artificial intelligence (AI) to create content for their followers.
The term AI covers a wide range of digital technologies, meaning everything from crunching huge amounts of digital data for analysis to what's known as “generative AI,” which “generates” new text or visuals based on large amounts of data. That's how this song by Peter Griffin was created.
“The rapid democratization of AI-generated technologies in recent years is having a major impact on how extremist groups exert their influence online,” Daniel Siegel, a US researcher who has analysed how AI is being used for malicious purposes, wrote in an article for the Global Network on Extremism and Technology. In it he featured a video of Peter Griffin.
AI-enabled extremism on the rise
Over the past year, various extremism monitors have reported that ISIS and other extremist groups are encouraging their followers to use new digital tools.
In February, an al-Qaeda-linked group announced it would be holding an online AI workshop. The Washington Post reportThe same group later published a guide on using AI chatbots.
After an Islamic State faction carried out a terror attack on a Moscow theater in March, killing more than 135 people, one of the group's supporters created a fake news show about the incident and released it four days after the attack.
Earlier this month, Spanish Interior Ministry officials Of the nine young people who were spreading propaganda praising IS across the country, one was allegedly focusing on “extremist multimedia content using a special editing application powered by artificial intelligence.”
“What we know now about the use of AI is that it acts as a complement to the official propaganda of both al-Qaida and ISIS,” said Mustafa Ayad, executive director for Africa, the Middle East and Asia at the London-based Institute for Strategic Dialogue, which researches all forms of extremism.
“This allows advocates and informal support groups to create emotive content that is used specifically to activate their supporter base around a core concept.”
He added that, judging by its appearance, it might not even be picked up by content moderators on popular social media platforms.
In fact, Ayad told DW that even though IS content is more absurd and unrealistic, it's often too novel to be shared among followers.
Extremists are “early adopters”
This comes as no surprise to longtime observers of the Islamic State group: When the militant group first rose to prominence in 2014, it was already producing expensive propaganda videos to intimidate its opponents and rally supporters.
“All of this speaks to what ISD has always pointed out,” Ayad said: “Terrorist groups and their supporters continue to be quick to adopt technology for their own gain.”
But how dangerous is this kind of content really? After all, the fake news broadcast about the Moscow attack looks fake and Peter Griffin's song isn't hurting anyone. But is it really?
Watchdog groups have listed a variety of ways extremist groups could use AI. Beyond propaganda, experts suggest that large-scale language model chatbots like ChatGPT could be used to talk to potential new members. If the chatbot piques interest, human recruiters could take over, experts say.
AI models like ChatGPT also have specific rules written into their systems to ensure they don’t help users with things like getting away with murder, but these rules have proven unreliable in the past, and it’s possible that a would-be terrorist could ignore them and obtain compromising information.
There are also concerns that extremists could use AI tools to carry out digital and cyber attacks, or even plan real-world terror attacks.
Deepfakes vs. real bombs
Experts say that while AI has terrifying potential in the hands of extremists, it is even more dangerous in real life.
In a 2019 paperWriting in the journal Perspectives on Terrorism, researchers looked at the relationship between the amount of propaganda disseminated by the Islamic State group and actual physical attacks, concluding that there was “no strong, predictable correlation.”
“This is similar to the discussion around 10 years ago about cyber weapons and cyber bombs,” said Lilly Pinenberg Mueller, a cyber security expert and research fellow at the School of War Studies at King's College London.
Today, even rumors and old videos, not to mention AI, can have a destabilizing effect and spark a flood of disinformation on social media, she told DW. “States have conventional bombs they can drop if that is their intent.”
“At this stage it is unclear whether the use of AI by foreign terrorist groups and their supporters is more dangerous than the very real and graphic propaganda they use to indiscriminately kill civilians and attack security forces,” ISD's Ayad said.
“The bigger threat now is that these groups will actually carry out attacks, inspire lone perpetrators, or successfully recruit new members out of the geopolitical situation, specifically Israel's response to the October 7 war on Gaza,” he continued. “They are using civilian deaths and Israeli actions as rhetorical tools for recruitment and campaign building.”
Edited by: Davis Van Opdolp
