Tackling AI Deepfakes in Politics: Google's Proactive Approach

• Deepfake
• AI Regulation

In response to the increasing popularity of artificial intelligence (AI) in political advertising, tech giant Google has decided to enforce transparency. Political ads that utilize AI must disclose when imagery or audio has been manipulated synthetically. The new policy is set to take effect in mid-November, according to The Associated Press.

A Global Policy Update

Google announced this political policy update to adopt similar policies for campaign ads in time for elections in the European Union, India, South Africa, and other regions where Google has a verification process. The policy comes as AI-generated and falsified media clips have become every day in political media. Generative AI tools make it easier and faster to produce misinformation, enabling bad actors to mimic speech or appearance in photos and videos more realistically.

The Impact of AI on Political Campaigns

AI-generated video has already been used by political campaigns, including one of the current forerunners for the Republican party in the US, Gov. Ron DeSantis of Florida. DeSantis’ campaign put out an ad that depicted his GOP opponent and Republican frontrunner, Donald Trump, positively embracing Dr. Anthony Fauci, a chief medical expert who advised Trump during the COVID pandemic. Similarly, the Republican National Committee (RNC) released a wholly AI-generated ad depicting the future under Joe Biden.

Federal Regulation of AI and Deepfakes

In line with Google’s new policies, the Federal Election Commission (FEC) is considering regulating AI-generated ads such as ‘deepfakes’. Advocates believe this could steer voters away from misinformation, as deepfakes can misrepresent political figures saying or doing things they never expressed in real life. Democratic senator Amy Klobuchar is a co-sponsor of legislation that would demand similar requirements to Google’s policy law; potentially deceptive AI-generated political ads will have to include disclaimers disclosing the fact.

Exemptions and Scope of the New Policy

It's important to note that this new policy does not mean all use of AI by political campaigns is banned. There are notable exceptions for altering content in ways that don’t change the substance and scope of the advert. For instance, this includes using AI tools for media editing and quality improvement purposes. The policy will broadly apply to YouTube and the rest of Google’s platforms and third-party sites within Google’s ad display network.

Other Tech Giants' Policies

As of this week, Google is still the only platform to put a policy like this in place in what is a proactive effort. Other social media platforms might have to follow suit if their approach needs to be revised, primarily if more widespread legislation exists. Meta, the parent company of Instagram and Facebook, doesn’t have an AI-specific prescriptive policy but a general blanket policy against “faked, manipulated or transformed” audio and imagery for misinformation purposes. Meanwhile, TikTok bans political ads altogether.

The Future of AI in Politics

The use of AI for political gains is still relatively uncharted territory. However, companies like Google are taking positive steps towards addressing the challenges associated with AI in political campaigning. This forward-thinking approach demonstrates a commitment to transparency and safeguarding the integrity of democratic processes. Other tech giants and legislators must follow suit and implement comparable measures.