New legislation from Democratic Sen. Annette Taddeo looks to regulate the use of “deepfake” technology when used in campaign materials.
The technology can be used as a sort of advanced CGI to make it look as though an individual said or did something they did not do or say.
So far, the emerging technology has mostly been used in obviously manipulated videos to demonstrate its potential. But in theory, videos could be crafted to fool viewers into thinking a candidate backed a controversial position or made an offensive statement, and the technology will only improve over time.
Taddeo is looking to stamp out that potential before it becomes mainstream. Her legislation (SB 658) targets any materials where a candidate’s likeness has been manipulated “with the intent to cause the viewer or listener of such medium to believe that a candidate is publishing a message that was not actually published by the candidate.”
Those materials must contain “a clear and conspicuous disclaimer,” according to the bill’s language.
“If the manipulated medium is an image or a video, the disclaimer must be displayed in text that is clearly visible to the average person for the duration of the publication,” Taddeo’s bill reads.
“If the manipulated medium is in audio form, the oral disclaimer must be made in a manner that is clear and intelligible to the average person and announced both at the beginning and the conclusion of the audio. If the manipulated medium in audio form is more than 30 seconds in duration, the oral disclaimer must run at the beginning and conclusion of the audio and also once every 20 seconds throughout the audio’s duration.”
More precise rules regarding those disclaimers will be crafted by the Division of Elections and the Consumer Protection Division of the Office of the Attorney General, if approved.
The first violation of the law would be a noncriminal infraction. Subsequent offenses would be classified as first degree misdemeanors.
Taddeo does not yet have a co-sponsor for companion legislation in the House.