Deepfake Detection and the Fight Over Robin Williams' AI Ghost

Robin Williams' daughter publicly slammed AI-generated versions of her late father, calling them disgusting. Someone built a synthetic Robin Williams, and his family had to watch a digital ghost perform without permission.

The dead cannot consent

Deceased celebrities are the easiest deepfake targets because they cannot object and their likeness data is everywhere: decades of films, interviews and appearances. Generative AI can now resurrect a performance convincingly enough to fool casual viewers. The result gets shared as novelty, or worse, monetised.

Families call it violation. Creators call it tribute. The law, in most places, has not decided what to call it. Meanwhile the technology keeps improving, and every beloved dead celebrity is one viral video away from an unauthorised comeback tour. This will get worse before it gets better. As generation quality improves, expect entire fake interviews and performances. Grief makes people vulnerable, and the fakers know it. The only durable defence is a habit: check before you believe, every single time. Her objection was public and blunt. The least we can do is verify before sharing. Share nothing unverified.

Deepfake detection is how the rest of us tell tribute from theft. An ai deepfake detection tool can flag a "new" Robin Williams clip as synthetic in seconds, which is also what tools help newsrooms verify whether media is ai-generated before they amplify it. Deepfake detection will not settle the ethics, but it settles the facts: this is not him. That distinction matters.

Seen a "new" video of someone who passed away? Try the AI Detector before sharing.