In May 2024, something unprecedented appeared on screens across Central Asia. A 52-second video in Pashto featured a news anchor calmly claiming responsibility for a terrorist attack in Bamiyan, Afghanistan. The anchor looked local, spoke fluently, and delivered the message with professional composure. There was just one problem: the anchor did not exist. The Islamic State Khorasan Province (ISKP) had produced its first AI-generated propaganda bulletin, and the implications for global security, content moderation, and the very architecture of our information ecosystem would prove profound. This was not an isolated experiment. Days later, ISKP released another AI-driven segment, this time featuring a synthetic anchor dressed in Western attire to claim responsibility for a bombing in Kandahar.…