AI use in political ads will require disclosure on Google and YouTube

Google will soon require that political ads using artificial intelligence be accompanied by a prominent disclosure if imagery or sounds have been synthetically altered.

AI-generated election ads on YouTube and other Google platforms that alter people or events must include a clear disclaimer where users are likely to notice it, the company said in an update this week to its political content policy.

The new rule comes into effect in mid-November, just under a year before the United States presidential election. It will also affect campaign ads ahead of next year’s elections in India, South Africa, the European Union and other regions where Google already has a verification process for election advertisers.

Though fake images, videos or audio clips are not new to political advertising, generative AI tools are making them easier to create and more realistic. Some presidential campaigns in the 2024 race, including that of Florida Governor Ron DeSantis, already are using the technology.

The Republican National Committee released an entirely AI-generated ad in April meant to show the future of the US if President Joe Biden were re-elected. It employed fake but realistic photos showing boarded-up shopfronts, armoured military patrols in the streets and waves of immigrants creating panic.

In June, DeSantis’ campaign shared an attack ad against his Republican primary opponent Donald Trump that used AI-generated images of the former president hugging infectious disease expert Anthony Fauci.

Last month the Federal Election Commission began a process to potentially regulate AI-generated deepfakes in political ads ahead of the 2024 election. Such deepfakes can include synthetic voices of political figures saying something they never said.

Democratic US Senator Amy Klobuchar, co-sponsor of pending legislation that would require disclaimers on deceptive AI-generated political ads, said Google’s announcement was a step in the right direction but “we can’t solely rely on voluntary commitments”.

Several states have discussed or passed legislation related to deepfake technology.

Google is not banning AI outright in political advertising. Exceptions include synthetic content altered or generated in a way that’s inconsequential to the claims made in the ad. AI can also be used in editing techniques such as image resizing, cropping, colour, defect correction or background edits.

The ban will apply to election ads on Googlown platforms, particularly YouTube, as well as third-party websites that are part of Google’s ad display network.

Google’s action could put pressure on other platforms to follow its lead. Facebook and Instagram parent Meta doesn’t have a rule specific to AI-generated political ads but already restricts “faked, manipulated or transformed” audio and imagery used for misinformation. TikTok doesn’t allow any political ads. X, formerly Twitter, didn’t immediately reply to an emailed request for comment.