Reporting and expert interviews show extremist groups adopting AI voice cloning and text‑to‑speech to produce multilingual propaganda. The Guardian highlights the use of cloned historical speeches and audiobooks to boost engagement and evade moderation.

A Guardian investigation details how extremist actors, including neo‑Nazi and jihadist networks, are rapidly adopting generative audio tools to produce more persuasive, multilingual propaganda. Sources and experts interviewed for the piece describe several misuse patterns: cloning voices of historical figures and movement leaders to create emotionally resonant speeches, producing narrated audiobooks of banned or extremist texts, and repurposing celebrity or influencer audio to lend credibility to messaging. The resulting content is optimized for platform formats — short, shareable clips and translated versions tailored to regional audiences — that increase reach and engagement across social networks, encrypted messaging apps and fringe forums. Moderation and regulatory responses are struggling to keep pace because synthetic audio can be produced quickly, distributed across decentralized channels, and frequently reuploaded with minor edits to evade automated detection. The article calls for coordinated action: investment in detection tools, clearer platform policies addressing synthesized voices, cross‑border cooperation among technology companies and regulators, and targeted counter‑messaging to blunt the appeal of emotionally powerful, fabricated audio used by violent and extremist movements.