Sudan's Battlefield Moved to the Feed

As Sudan's civil war ground on, researchers documented ethnic hate speech, dehumanizing propaganda and a flood of AI-fabricated video and audio across social platforms, some of it tied to real-world attacks. In October 2024, X suspended the accounts of the Rapid Support Forces and its commander amid the campaign.

Since fighting between the Sudanese Armed Forces and the paramilitary Rapid Support Forces erupted in April 2023, a second front has opened on social media. In an analysis for the Tahrir Institute for Middle East Policy, researcher Hamid Khalafallah wrote that the conflict has been accompanied by a “virtual propaganda warzone” in which disinformation and fabricated news are deployed as a systematic strategy of psychological warfare, by both sides.

The stakes are not only reputational. A Thomson Foundation report by researchers Aida Al-Kaisy and Amal Hamdan, cited by Forbes, described Sudan's information ecosystem as “a battlefield as brutal as the physical conflict itself” and found that disinformation serves “as a powerful mechanism to ensure the war continues.”

Fabricated voices, real confusion

Among the earliest documented cases of synthetic media in the war was a campaign impersonating Sudan's former president, Omar al-Bashir. Beginning in late August 2023, a TikTok channel calling itself “The Voice of Sudan” posted purported leaked recordings of al-Bashir that drew hundreds of thousands of views, according to a BBC investigation reported by African Arguments. Analysts who examined the audio for the BBC concluded the voice had been generated with AI voice-cloning tools.

Hany Farid, a digital-forensics professor at the University of California, Berkeley, who analyzed the recordings for the broadcaster, said the broader danger lay in how accessible the technology had become. “It is the democratisation of access to sophisticated audio and video manipulation technology that has me most worried,” Farid said, according to UC Berkeley.

Later fabrications were tied more directly to violence. African Arguments reported that in March 2024 an AI-generated audio recording, attributed to the head of the Sudanese Armed Forces and appearing to order the killing of civilians, was viewed some 230,000 times and shared by hundreds of accounts, including those of known Sudanese politicians.

Hate speech that precedes attacks

Researchers have also traced a path from online incitement to physical harm. In one case cited in the Thomson Foundation report, a Sudanese Armed Forces officer posted on Facebook accusing residents of the Khartoum-area district of Shambat of collaborating with the Rapid Support Forces; within two days, a community kitchen there was struck by a suicide drone, Forbes reported.

The volume of synthetic material rose sharply as the war widened. After the Rapid Support Forces seized the western city of El Fasher in October 2025, an interviewee told the Thomson Foundation researchers that an estimated 90 percent of the videos and posters that spread from the city were AI-generated rather than authentic, a flood that let combatants deny killings and blur genuine evidence, according to Forbes.

Some researchers caution that the fabrications are, for now, often crude. Mohamed Sabry, a Sudanese AI researcher at Dublin City University, said many of the deepfakes are “easily identified as fake because of their poor quality,” which he attributed to a lack of training data in Sudanese dialects, African Arguments reported. He warned that the weakness would fade as the tools improve.

Platforms move, unevenly

The campaigns have drawn intermittent enforcement. In October 2024, X suspended the official account of the Rapid Support Forces along with the account of its commander, Lt. Gen. Mohamed Hamdan Dagalo, known as Hemedti, and several linked commanders, as reported by Dabanga. The platform said the accounts had violated its rules by promoting violence and human-rights violations and by spreading hate speech. Facebook and YouTube had taken similar action against RSF accounts earlier in the conflict.

Investigators say the manipulation extends to the mechanics of the platforms themselves. In its analysis for the Tahrir Institute, TIMEP reported that the Rapid Support Forces have been accused of hijacking at least 900 dormant X accounts to perpetuate their propaganda. Drawing on the UK-based digital agency Valent Projects, which examined tweets using pro-army hashtags, the institute also reported that only about 10 percent of the examined tweets appeared to come from authentic accounts, the institute wrote.