How deepfakes, AI-generated videos that graft a person's face onto another's body, have been weaponized to harass and humiliate their subjects, mostly women
“Deepfake” creators are making disturbingly realistic, computer-generated videos with photos taken from the Web, and ordinary women are suffering the damage.
Context & Ripple Effects
This late-2018 Washington Post report was the early warning shot: face-swapping tools had moved from celebrity parody to a harassment weapon aimed at ordinary women, with nothing more than web-scraped photos as raw material. At the time there was essentially no legal architecture for the harm.
The coverage that followed shows how that vacuum got filled piecemeal. By mid-2023, state laws in Virginia and California criminalized sharing nonconsensual deepfake porn, but advocates were still calling for a federal statute; by late 2023, lawmakers were scrambling over AI-generated nudes of teen girls; and researchers at DARPA had begun work on detection after warnings that convincing fakes would give cover to dismissing real events.
First-order effects
- Ordinary women whose photos appear online become involuntary raw material: the direct harm lands on private individuals, not just public figures, and the videos spread faster than takedown processes can respond.
- Platforms and communities hosting face-swap tooling and the resulting videos face immediate pressure to define and enforce a consent standard they did not previously have.
Second-order effects
- Legislatures are forced into a patchwork response — Virginia and California moved first on sharing bans, leaving a gap that only federal legislation could close, which is exactly the fight the later coverage documents.
- The credibility damage spreads beyond victims: once fakes are this convincing, accused wrongdoers gain a ready-made defense that real footage is fabricated, pushing funders like DARPA into detection research.
Third-order effects
- If the pattern holds, synthetic media forces a durable legal category — likeness governance — where a person's face and body become protected assets requiring consent infrastructure, much like data privacy became one.
- Detection and provenance tooling shift from academic curiosity to required platform plumbing, because enforcement against fakes is impossible without a way to authenticate real media.
The trend: Synthetic media is completing its move from novelty to weaponized harassment, dragging law, detection research, and platform policy toward a consent-based regime for digital likenesses.