TAKE IT DOWN Act becomes law, creating national framework for non-consensual AI imagery removal
In brief
- President Trump signed the TAKE IT DOWN Act into law on May 19, 2025
- Senate passed unanimously in February 2025; House cleared it 409-2 in April
- Platforms must remove non-consensual intimate imagery after receiving valid takedown notices
- Companion DEFIANCE Act empowers victims to file civil lawsuits against deepfake creators
- First conviction under the TAKE IT DOWN Act occurred in April 2026
Bipartisan Momentum
Senator Ted Cruz and Senator Amy Klobuchar introduced the bill on January 16, 2025. The TAKE IT DOWN Act passed the Senate by unanimous consent in February 2025, then cleared the House with a 409-2 vote on April 28, 2025. The speed was striking. Congress found something it could actually agree on: AI-generated fake intimate images are bad.
First Lady Melania Trump also publicly championed the bill, lending high-profile momentum to the effort. That backing, combined with bipartisan sponsorship, helped the legislation move faster than most tech bills do.
The Law and Its Teeth
The TAKE IT DOWN Act creates a national framework requiring platforms to remove non-consensual intimate imagery after receiving a valid takedown notice. Platforms have a clear, enforceable obligation. No ambiguity about whether the content violates their terms. The takedown notice itself is the trigger.
A companion bill, the DEFIANCE Act, advanced through the Senate by unanimous consent in January 2026. This one empowers victims to file civil lawsuits against the creators and distributors of deepfake content. Criminal penalties and civil recourse. Together, the two laws create a two-front deterrent.
Early Enforcement
The first conviction under the TAKE IT DOWN Act came in April 2026, involving an individual in Ohio who used AI to create and distribute non-consensual imagery. It's early, but enforcement is happening. The law is moving from the statute book into courtrooms.
The TAKE IT DOWN Act is one of the fastest-moving pieces of AI legislation in recent memory. The speed signals that Congress can act decisively on AI harms when the harm is clear and the damage is personal. Whether that pattern holds for other AI risks—or whether deepfake legislation remains the exception—remains an open question.
