This browser does not support the Video element.
AI generated political campaign ad draws controversy
We're living in an era where artificial intelligence is being used for more areas of everyday life, and now it's also becoming a tool in politics. One campaign ad is stirring up controversy in what some view as an AI-generated attack ad against Lt. Governor Peggy Flanagan, who is running for U.S. Senate. FOX 9's Soyoung Kim has more.
MINNEAPOLIS (FOX 9) - Artificial intelligence (AI) is a tool used in many functions as part of everyday life and is also becoming a tool used in politics.
One campaign ad stirring up controversy is what some view as an AI-generated deepfake attack ad against Lt. Governor Peggy Flanagan, who is running for U.S. Senate.
Experts say this raises some questions that have yet to be tested.
Crossing the line?
The backstory:
Lt. Governor Peggy Flanagan is the Minnesota DFL's endorsed choice to run for U.S. Senate. This comes after Democratic U.S. House Representative Angie Craig announced she will not seek the party's endorsement.
An attack ad created by a super PAC supporting Craig targeting Flanagan is at the center of controversy. Dozens of DFL legislators have spoken out condemning the use of AI-generated deepfakes in political advertising.
What they're saying:
FOX 9 spoke with Manjeet Rege, the Director of the Center for Applied Artificial Intelligence and Professor and Chair of the Department of Software Engineering and Data Science at the University of St. Thomas, who helped break down some tips to look out for to spot a deepfake video. Rege says look for duplicates in the background or rough edges on objects.
"If I would pick two things, it would be the gaze. The gaze is the same, it is not moving. I’m also looking a little below the camera, not exactly at the camera. And the other is the speed, the rate at which I’m speaking and my tone of speaking," said Rege, explaining some ways to detect a manipulated deepfake video by using an example he had created of himself.
Rege explained how he created multiple deepfake versions using his own image speaking languages he does not know.
"When you take deep learning technology, apply that to image or video data and create content. The content that gets created is fake, looks realistic, but created using deep learning, hence it is called deepfakes," said Rege.
"All of these examples you’re seeing are my deepfake versions. In some of them, I’m speaking languages I do not speak like I’m speaking Chinese and French in some. They ask you to record 30 seconds. Then based on that 30 seconds it is going to then extrapolate that to how would that person react to different situations."
Does the content violate state law?
What they're saying:
FOX 9 political analyst Blois Olson says the questions to ask are does the content violate state law that applies within 90 days of an election.
"In 2023, Minnesota Legislature passed a law that made deepfakes in political communications and ads illegal. The question is what’s a deepfake, how do you define this. It has not been tested in the courts," said Olson.
The Minnesota law sets the threshold at whether a reasonable person would believe it as reality. Olson says AI in politics is new territory, but says what is not new are exaggerations in politics.
"I had seen the ad and then I saw it being accused of being a deepfake, and then I had to go back and rewatch it," said Olson.
"I watched it 15-20 times by now and I still look at the entirety of the ad and what it’s saying to know that the vast majority of this ad is a very traditional political attack ad on somebody," said Olson.
What’s next: Flanagan’s campaign said the ad misleading and said they are considering legal options, but did not expand on details they are discussing or exploring.
Craig’s campaign said the ad in question is not from their campaign, and added that the Congresswoman does not support the use of AI in political ads.
What's next:
Both DFLers are running for U.S. Senate in the August primary.