Personal & Political Sovereignty

AI lowers the cost of generating convincing political content — deepfakes, microtargeted messaging, synthetic personas — to near zero. The defenses of democratic discourse (shared factual baseline, identifiable speakers, slow media cycles) have not adapted. Voters, citizens, and consumers increasingly cannot trust what they see.

Self-governance depends on the capacity to perceive reality accurately enough to make informed choices. When perception itself is industrially manipulated, the choice that follows is not autonomous. The 2026 midterms are the first US election cycle where AI-generated political content is mainstream.

30 states have enacted election deepfake laws; federal action remains stalled. EU AI Act (effective August 2026) and Italy's national AI law are global firsts. C2PA provenance adoption is accelerating but platforms still strip metadata on upload.

AI enables information operations at scale; algorithmic curation shapes political reality; deepfakes erode truth baseline; surveillance capitalism + state surveillance harvest data without sovereignty; IP/data laws not aligned with individual property rights. The market produces personalized epistemic environments rather than shared ones.

Platform companies (engagement), state surveillance apparatus, political operatives (microtargeting), advertisers, foreign IO actors. Polarization is itself a business model.

Section 230 reform — stalled. Fact-checking — limited reach, polarized reception. Content moderation — whack-a-mole. FTC privacy actions — under-resourced. "Media literacy" curricula — slow, contested.