A digital safety coalition told Newsweek that artificial intelligence (AI) has enhanced content for lone-wolf terrorist attacks by ISIS and Al-Qaeda against Jewish communities in the United States.
"The trend is unmistakable," Marc Ginsberg, a former U.S. ambassador and the president of the Coalition for a Safer Web, told Newsweek. "That is a whole new level of threat that I don't think anybody understands or appreciates how sophisticated the penetration operation has really become."
Newsweek reached out to the Department of Homeland Security via email for comment.
Why It Matters
In the year following the October 7, 2023, attack by Palestinian militant group Hamas in Israel, the Anti-Defamation League (ADL) has counted more than 10,000 antisemitic incidents in the U.S. This is the highest number of incidents ever recorded in a single year since the ADL started tracking in 1979, representing an over 200 percent increase compared to the incidents reported during the same year prior.
Attacks have included verbal and written harassment, vandalism and physical assault. At least 1,200 incidents between the 2023 attack and its 2024 anniversary were on college campuses.
Following the anniversary of October 7, the Coalition for a Safer Web found numerous ISIS and Al-Qaeda websites accompanied by generative AI videos, programs and memes of Gaza's destruction and injured Palestinians. The content is manipulated to appeal to potential lone-wolf recruits in the U.S. to start retribution against Israel's supporters in the U.S., particularly the Jewish community.
What To Know
Since mid-2024, ISIS and Al-Qaeda have been upgrading their propaganda by incorporating generative AI programs, according to Ginsberg and his team's research. The AI programs create a "super stylized" version of extremist web content. It also allows terrorist groups to manipulate algorithms and breach social media content, inciting antisemitic operations.
"Generative AI has been hijacked by ISIS and Al-Qaeda in their respective campaigns to achieve global dominance over the other, to leverage the Gaza war, to incite and recruit domestic based U.S. nationals and not to attempt to infiltrate any personnel actually from their operations into the U.S.," Ginsberg told Newsweek.
The Coalition for a Safer Web found "help wanted" ads to recruit AI software developers and producers. The generative AI creates news programs in an Americanized English sound clone, pushing content geared toward a younger, more persuadable demographic in the hopes that they would incite a potential terrorist attack.
The ISIS and Al-Qaeda sites can fly under the radar due to the generative AI since it obscures their identities.
The coalition found several how-to guides and tips on how to use cyber propaganda in publications like The Wolves of Manhattan magazine, Mujahideen in the West, and the Voice of Khurasan.
Ginsberg noted that there are "honey trap" social media platforms, particularly TikTok, Instagram, X (formerly Twitter) and YouTube, that have ended content moderation efforts. This has, in turn, allowed ISIS and Al-Qaeda to upload antisemitic content more easily.
"I think that when you introduce any new software capacity, whether it be encryption, whether it be conversely a termination of social media moderation on major platforms, the terrorists are going to find every way which possible to try to beat the system that is leveraged against them," Ginsberg told Newsweek. "What's unique about this is that the content that we're intercepting doesn't have breadcrumbs enabling us to track this back directly to a particular source."
The coalition also found "target identification packages," which contain several videos of photos of Jewish centers in New York, Miami, Chicago and Detroit, as well as Sydney, Melbourne and Toronto.
The Growth of AI Leaves The Opportunity For This To Get Worse
Ginsberg said that "of course," as AI continues to grow, the risk of additional threats also increases "because it's becoming uncontrollable."
"Until the last few months, we were able to go onto most encrypted and unencrypted platforms that are used by mainstream young Americans," Ginsberg told Newsweek. "The threat level is something that's only getting worse. Everybody, including the ADL, has said the same, but what is anybody doing about it?"
Clara Broekaert, an analyst at the Soufan Group who studies new and emerging technologies with open-source intelligence investigations, told Newsweek that generative AI needs to fulfill a certain need for terrorist groups.
"Very early on," Broekaert said, her group was able to see that generative AI and the creation of synthetic content was very quickly adapted experimentally by extremist actors, particularly far-right, neo-Nazi and government exploration groups.
"It's enormously beneficial for mass translating propaganda material to the vast community that they have online, stretching all the way from Central Asia where so many languages are spoken to the West to the South to Somalia to Mozambique," Broekaert told Newsweek. "This was really important in their propaganda operations, facilitating them and centralizing them to some extent."
Broekaert agreed with the Coalition for a Safer Web that generative AI makes propaganda less detectable. However, she added that it's "not a huge disrupter in how terrorists think about their propaganda."
"They push out online incitements to commit attacks, and many of them don't have any AI component to it, so I don't necessarily think this will have a direct impact," Broekaert told Newsweek. "But I can completely imagine that once we have something like very robust predictive analytics, this may help terrorist organizations plan efficiently."
In general, AI assists with gathering information through large language models (LLM). For example, Broekaert brought up how the attack at the Trump International Hotel in Las Vegas on New Year's Day had collected guides through LLMs.
This makes it harder to combat such an issue because AI, like ChatGPT and other proprietary models that are not open source, is not necessarily fine-tuned. Broekaert suggested there needs to be guidelines to ensure they don't give out harmful information regarding how to plot an attack.
A Surge In Terror Plots Against U.S. Jewish Targets Since Late 2024
The FBI and Canadian authorities arrested 20-year-old Muhammad Shazeb Khan, a Pakistani citizen, while he was en route to a planned mass shooting in support of ISIS at a Jewish center in Brooklyn on September 4, 2024. The shooting would have coincided with the anniversary of the Hamas attack on Israel.
On October 8, federal agents in Oklahoma City arrested 27-year-old Afghanistan national Nasir Ahmad Tawhedi, who was reportedly incited by ISIS online operatives to execute a terror attack on Election Day. Propaganda was saved to Tawhedi's iCloud and Google accounts. He also participated in pro-ISIS Telegram groups.
Just 10 days later, an Arizona teen, Marvin Aneer Jalo, was arrested and charged as an adult for plotting an ISIS-inspired drone attack on the Phoenix Pride Festival parade, potentially as it passed by a Jewish community center, according to the coalition.
On October 26, police arrested 22-year-old Sidi Mohammed Abdullahi in Chicago for shooting a Jewish man on his way to synagogue. Abdullahi turned his weapon on the first responders.
Houston FBI Counterterrorism Task Force arrested Anas Said on November 8. The 28-year-old Houston native was taken into custody for conspiring with ISIS web designers to plot a mass casualty attack against a Houston Jewish installation.
On December 18, Abdullah Ezzeldin Taha Mohamed Hassan, an 18-year-old student at George Mason University in Virginia, was arrested in connection with a plot inspired by ISIS content on X to stage a mass attack against the Israeli Consulate in New York City.
He was going to livestream the attack and create a "martyrdom video," according to the coalition.
What People Are Saying
Michael S. Smith II, a terrorism analyst who specializes in the influence operations of al-Qaeda and ISIS, told Newsweek: "I believe that, just as we have seen with the weaponization of social media, it is reasonable to assume ISIS, al-Qa'ida and a wide range of terrorist organizations and extremist elements, as well as a long list of intelligence agencies, will come to view AI as a tool that offers extraordinary potential to do deep damage to global security.
"And while I believe the posture against this will be largely reactive on the parts of various global powers, I believe the reactions will be more forceful than the responses to al-Qa'ida's, followed by ISIS's uses of social media to widen their capabilities to threaten the U.S. and its closest allies."
Halie Soifer, CEO of the Jewish Democratic Council of America, sent Newsweek a press release following President Donald Trump's talk with Prime Minister Benjamin Netanyahu on Tuesday: "Jewish Americans overwhelmingly reject and disapprove of Donald Trump because his policies and rhetoric are antithetical to our values. Tonight's press conference with Prime Minister Netanyahu was no exception."
What Happens Next
The Coalition for a Safer Web laid out several measures that the public and private sectors can take to neutralize and confront the antisemitic terrorist groups.
Primarily, Ginsberg suggested that the White House needs to issue a new antisemitic executive order. On January 29, President Donald Trump signed an executive order to add additional measures to combat antisemitic harassment and violence. Ginsberg believes, however, that this or other federal resources should explicitly target cyber content moderation.
Along those lines, Ginsberg also added that the definition of terrorist "material support" should be amended under U.S. laws to include technology and software support enabling extremist and terrorist groups to incite domestic terror.
As far as the coalition can tell, the U.S. code does not explicitly include or reference digital tools in its definition—just "tangible or intangible...communications equipment."
"If there is a will to develop counter-targeting in order to neutralize and to drop what is essentially poison pills through AI into these efforts and to have more of a holistic approach, it's an essential component of mitigating the threat of domestic terrorism initiatives," Ginsberg told Newsweek.
English (US) ·