AI-Enabled Disinformation and Geopolitical Influence

China launches 'Data Center Bandwagon' campaign using ChatGPT-generated comics to oppose US data center construction, exploiting local opposition. New tool (SAGA) traces fake videos to source using temporal attention signatures. Subtle LLM bias research reveals narrative shaping through creative homogenization, opinion steering, and historical bias. This emerging threat combines AI-generated content with geopolitical influence operations. Latest: KAIST research uses 20 years of Naver comments to detect foreign-linked opinion manipulation with 23,998 suspicious accounts found, a practical disinformation defense. Tech's push for AI content traceability highlights multi-layered provenance (watermarks, hardware signing, metadata) but acknowledges no silver bullet. AI detectors technical guide provides foundational knowledge on detection mechanics.