Face reenactment, also called puppeteering, drives a target’s facial expressions and head movement using a separate performer’s input. Unlike face swap, the target’s actual face is preserved; only the expressions and timing are controlled by the attacker.

• A driving video of the attacker is captured in real time or pre-recorded
• A model maps the attacker’s expressions onto the target’s face while preserving the target’s appearance
• The target appears to say or do things the attacker performs, with natural facial dynamics

• Manipulated political speeches where a real candidate is made to deliver scripted content
• Fabricated executive statements pushed to media or investors
• Fake interviews and testimonials using public figures
• Targeted disinformation aimed at specific demographics or regions
• Synthetic content for satire that crosses into defamation or election interference