AI Gets Political: How Can We Prevent Fake News from Election Ads?

Deceptive political advertising is nothing new. But as one news outlet recently reported, generative artificial intelligence (AI) is ushering in “a new era that could make it even harder for voters to separate truth from lies.”

A Wall Street Journal article dated June 5 agreed, noting that sophisticated AI-generated videos and images pose a “great threat to political campaigns as the 2024 election campaign begins.” claimed.

For example, in late May, Forbes reported that the super PAC, which backs Republican presidential candidate and Florida governor Ron DeSantis, used AI to create a video that falsely portrayed a “group of fighter jets emerging.” .[ing] It flies overhead as DeSantis speaks to the crowd. Alex Thompson debunked Axios’ fake flyover, calling it “the latest example of political advertising, including videos digitally altered to promote or attack candidates, and that viewers makes it difficult to discern what is true.”

So what can be done to combat political deepfakes?

There is no silver bullet. But the polarized journalism environment includes the seemingly right-leaning The Wall Street Journal and Fox News Channel, the seeming left-leaning The New York Times and MSNBC, and perhaps the more midstream legacy broadcast networks (ABC, CBS, NBC) will unite on this one important issue.

Each should assign journalists to dedicated inter-agency teams to work together to hold candidates of all political stances jointly accountable in courts of public opinion for deceptive AI-generated images and videos. Importantly, the side benefit of such an integrated and cooperative front here is that it restores a little bit of public trust in journalism and encourages people to move out of the silos of news sources and into broader journalism. It means that you may be able to gain the trust of

But the typical reaction when new technical problems arise is to adopt legislation. In fact, last month, Rep. Yvette Clarke (DN.Y.) amended the Federal Election Campaign Act of 1971 to require “clear and conspicuous” disclosures (called disclaimers) in campaign ads containing “images or video footage.” submitted a bill to mandate the Generated in whole or in part using artificial intelligence. “

Known as the REAL Political Advertising Act (the first part is a harsh acronym for ‘Call for AI-Driven Disclosure’), the bill would ‘explain the potential of AI-generated content to manipulate and deceive people on a large scale’. If there is, it could violate the Political Advertising Act,” Clark said. It could have devastating consequences for our national security and election security. “

Disclosures and disclaimers are certainly measures to deal with deceptive AI-generated campaign advertising, but they are likely to run into constitutional problems under other laws. That’s because the First Amendment to the U.S. Constitution provides a powerful bulwark against governments that prohibit false advertising of candidates. As First Amendment attorney Lata Knott wrote candidly in his 2020, “Lies in political advertising […] Part of the reason is that “political advertising is considered political speech, and the First Amendment protects political speech more than any other kind of speech.”

Furthermore, the U.S. Supreme Court explained in 2012 that it “never upheld a categorical rule.” […] False statements are not protected by the First Amendment. The court reasoned that for false speech to be exempt from First Amendment protection, it must cause some “legally justifiable harm,” such as defamation causing defamation.

False campaign advertising can be detrimental to democracy, but it justifies punishing false speech, such as financial harm to consumers who are lied to by salespeople to buy products. not a specific and easily identifiable harm such as Items that the salesperson knew were defective and were broken after purchase.

In addition, federal law intentionally and repeatedly denies radiobroadcast television stations “reasonable access” to candidates running for legally qualified federal office to air election advertisements. prohibited. Additionally, the Federal Communications Commission typically “reviews and pre-approves the content of political advertisements before they are aired” and “guarantees the accuracy of statements made by candidates and advertisers.” explains that it will not.

What can journalists do? One is fact-checking, which already exists. For example, the Florida-based Poynter Institute (owner of the Tampa Bay Times) is home to Politifact, which Poynter describes as “the largest political fact-checking news agency in America.” In the AI ​​age, fact-checking with visual images is just as important as verifying the verbal claims of candidates.

My proposal goes beyond media outlets doing their own siled fact-checking in three ways.

First, a broad coalition of organizations (discussed above) will target all legally eligible candidates in major election campaigns with ads that use AI to generate images and information that can be reasonably deceptive. will draft a statement asking them to jointly sign a declaration not to publish it on viewers or readers.

Second, all organizations will continue to update and prominently repeat the names of all candidates who refuse to sign and those who have signed.

Third, a fact-checking team of journalists from all groups worked together to decipher the deceptive visuals, and each group then reported the same co-authored article to establish the agreed findings. will be publicly exposed. The article will highlight the names of all news outlets and highlight their agreement with the findings.

Voluntary collaborative journalism efforts like this are not a panacea, even with all parties involved. Ultimately, it is up to voters to hold the deceitful ballot box candidates accountable. But such an approach by the United Press to this pressing political issue may help concerned voters while increasing credibility in journalism as a whole.

Clay Calvert, J.D., Ph.D., Professor Emeritus of the University of Florida (UF), Professor of Law at the Frederick G. Levin College of Law, and Brechner Distinguished Scholar in Mass Communications at the College of Journalism and Communication. Mr. Calvert specializes in First Amendment and media law. Non-Resident Senior Fellow At the American Enterprise Institute.