Tareq Saleh voice deepfake — Sep 2026

On September 13, 2026, reports emerged that Iran-backed Houthi fighters allegedly deployed artificial intelligence to mimic the voice of Yemeni commander Lt. Gen. Tareq Saleh. The incident occurred during a significant Houthi offensive near the strategic Red Sea port of Mocha, where the forces led by Saleh were experiencing mounting military pressure.

What happened

An audio recording began circulating among troops which appeared to feature the voice of Tareq Saleh, who leads the UAE-backed National Resistance Forces. The content of the clip featured the commander instructing his units to withdraw toward Mocha, framing the maneuver as a routine aspect of combat operations. The dissemination of this recording coincided with a broader Houthi push through government lines along Yemen’s western coast.

Shortly after the audio began to spread, the military media office for Saleh formally dismissed the clip as fabricated. Officials stated that the content was AI-generated and intended to sow chaos, confuse government troops, and undermine morale during active combat. In addition to the audio, reports indicated that messages were circulating in the region urging fighters to surrender their weapons and return home in exchange for amnesty.

Saleh’s forces have since cautioned that any statements or orders attributed to the commander that do not originate from official National Resistance military channels should be treated as unverified. While the military has labeled the audio as a deliberate act of psychological warfare, independent verification of the AI-fabrication claims remains pending, and the extent of the impact on troop morale has not been confirmed.