Ricketts Introduces Bill to Combat Deepfakes, Require Watermarks on A.I.-Generated Content

September 12, 2023

WASHINGTON, D.C. – Today, U.S. Senator Pete Ricketts (R-NE) introduced the Advisory for A.I.-Generated Content Act to require a digital watermark on materials generated by Artificial Intelligence (A.I.). It would direct the Federal Trade Commission (FTC), the Federal Communications Commission (FCC), the Department of Justice (DOJ), and the Department of Homeland Security (DHS) to establish federal guidelines for enforcing the watermark requirement, including on political advertising.

“With Americans consuming more media than ever before, the threat of weaponized disinformation confusing and dividing Americans is real,” said Senator Ricketts. “Deepfakes generated by artificial intelligence can ruin lives, impact markets, and even influence elections. We must take these threats seriously. My bill requiring a watermark on A.I.-generated content would give Americans a tool to understand what is real and what is made-up.”

Ricketts’ legislation requires the FTC, FCC, DOJ, and DHS to work together to establish guidelines for a required digital watermark on a variety of A.I.-generated materials for profit. The A.I.-generated materials include those impersonating someone else and products represented as original materials such as artwork, songs, and news. Upon the establishment of guidelines, A.I. generating companies would have one year to comply. The FTC would have primary enforcement authority.

Bill text can be found here.

BACKGROUND:

Generative A.I. tools have become popular for their ability to create original text and images when prompted. However, A.I.’s power to produce convincing text and photorealistic images has already been used to disseminate false information. In May, markets dropped briefly after a fake image of the Pentagon shrouded in smoke circulated on social media. The picture contained many unrealistic elements that occasionally crop up in A.I.-generated images, such as physical objects blending into each other.

A study published in June found that a majority of people were unable to tell whether a tweet was written by a human or by ChatGPT. The participants surveyed even found ChatGPT-tweets more convincing than their human counterparts.