Under the new policy advertisers must disclose whenever AI is used for political or social issue ads that feature photorealistic images or videos or realistic sounding audio that’s digitally created or altered but presented as real.

That includes ads that depict a real person as saying or doing something they did not say or do. Any ad using AI or digital creation that depicts a realistic-looking person that does not exist or a realistic-looking event that did not happen also must be disclosed.

The new Meta policy on AI use in political and social issue ads also requires disclosure when depicting a realistic event “that allegedly occurred, but that is not a true image, video, or audio recording of the event.”

Advertisers will have to complete an authorization process and include a “paid for by” disclaimer.

“If we determine that an advertiser doesn’t disclose as required, we will reject the ad and repeated failure to disclose may result in penalties against the advertiser,” Meta’s statement said. “We will share additional details about the specific process advertisers will go through during the ad creation process.”

Ads using AI or digital creation in ways that are not inconsequential or immaterial to the claim, assertion or issue raised in the ad would not have to be disclosed.

“In the New Year, advertisers who run ads about social issues, elections & politics with Meta will have to disclose if image or sound has been created or altered digitally, including with AI, to show real people doing or saying things they haven’t done or said,” Facebook President of Global Affairs Nick Clegg wrote on X.

In 2021 Meta changed some advertising practices, ending the advertising targeting of users based on interactions with “sensitive” matters like politics, race and sexual orientation.