Google is laying the ground rules for the use of artificial intelligence when it comes to political advertisements.
The tech juggernaut will require any political ads that use AI to create altered images or sounds to have a “clear and conspicuous” disclosure in a spot on the ad so that viewers will be able to see the warning.
Google even suggested wording that the company would find acceptable: “This image does not depict real events” and “This video content was synthetically generated,” BBC News reported.
The new rules were developed in response to the “growing prevalence of tools that produce synthetic content,” a Google spokesperson told BBC News.
The disclosures are similar to those that Google already requires indicating who paid for the ad.
The guidance goes into effect in November, a year before the presidential election. But there has already been synthetically created content being used in the campaign. The Washington Post reported that Florida Gov. Ron DeSantis’s campaign released a video showing former President Donald Trump hugging his former White House COVID-19 advisor Anthony Fauci.
Another ad showed Trump being arrested which was created by AI, BBC reported.
The false narrative isn’t unique to the U.S. A Polish opposition party used AI-generated audio to copy the voice of Prime Minister Mateusz Morawiecki, the newspaper and the European Federation of Journalists reported. Another deepfake clip showed Ukrainian President Volodymyr Zelensky considering surrendering to Russia.
Google is not banning the use of AI in advertising and will allow altered or generated synthetic content that has no impact on the ad itself. It can also be used in editing such as color correction, resizing and background editing, The Associated Press reported.
Google itself owns a “generative” AI tool called Bard. There is also OpenAI’s Dall-E that creates images, The Washington Post reported.
The rules only affect political advertisements, not regular videos on YouTube, The Washington Post reported.
Google is the first tech company to require an AI disclosure, Politico reported. The Federal Election Commission has yet to regulate if and when AI can be used in campaigns, but in August it agreed to seek public comment on whether to make updates to its policies to include AI advertisements.