Embodied Deepfakes in Public

Advances in neural lip‑syncing and soft humanoid hardware make it feasible to produce physically present robots whose mouth and facial motions closely match voiced audio, across languages. Such embodied deepfakes can be used for benign purposes (therapy, accessibility, entertainment) but also for impersonation, political spectacle, or covert influence in public spaces. — This shifts the deepfake debate from media provenance and content takedowns to in‑person identity, consent, public‑space signage, authentication, and criminal liability for impersonation or coordinated manipulation.