Newt Gingrich and Taylor Barkley Examine AI-Generated Political Ads, Voter Behavior, and the Future of Free Speech
In 2023, technology experts warned that artificial intelligence could make 2024 the last election truly decided by human voters. Three years and two election cycles later, that prediction has yet to materialize. Despite the growing use of AI-generated political advertisements, deepfakes, and sophisticated digital campaigns, voters continue to make decisions that defy the expectations of political consultants and technology experts. On this episode of Newt’s World, Newt Gingrich welcomes Taylor Barkley, Director of Federal Government Affairs at the Abundance Institute, to examine what artificial intelligence actually accomplished during the 2026 midterm elections and why the public may be far less susceptible to digital manipulation than many predicted.
Their conversation explores the surprising limits of AI-driven political advertising, the importance of trusted relationships in shaping public opinion, and the constitutional questions surrounding efforts to regulate AI-generated content. Barkley argues that existing laws already provide important protections against election deception and warns that overly broad restrictions could threaten political satire, innovation, and First Amendment rights. Gingrich and Barkley also consider what a new Congress could mean for artificial intelligence policy and why technological progress offers enormous opportunities for America’s future. Their discussion raises a fundamental question: In the effort to protect democracy from artificial intelligence, could government regulation become a greater threat to free expression than the technology itself?
Listen to their conversation below, or scroll down for an edited transcript.
Edited Transcript
This conversation has been condensed and lightly edited for clarity.
Newt Gingrich
Taylor, you recently published an op-ed that opens with a 2023 warning that “2024 will be the last human election.” Who made that prediction, and what did they mean?
Taylor Barkley
That quote comes from a video called The AI Dilemma, published by the Center for Humane Technology in 2023. The speakers were Tristan Harris and Aza Raskin. Raskin argued that whoever had the most computing power and the most sophisticated AI tools would ultimately win elections.
I’ve been tracking predictions about artificial intelligence and elections since 2019. Over and over, we’ve heard warnings that AI would fundamentally change how people vote, that the public would be deceived by synthetic media, and that political outcomes would be determined by technology rather than voters.
But we’ve now been through the 2024 election and the 2026 midterms. Humans are still in charge. The prediction was wrong.
Newt Gingrich
Walk us through what they got wrong. What did the actual election results tell us?
Taylor Barkley
If the prediction were correct, the party spending the most money on AI-generated political advertising should have won.
Republicans appear to have spent substantially more on AI-generated advertisements than Democrats during the 2026 midterms. Yet Republicans lost.
The public is much more resilient than pundits and technology commentators give them credit for. People aren’t automatically persuaded by a sophisticated video simply because artificial intelligence produced it.
That’s the real story. We should hold people accountable for predictions that create unnecessary panic about technology.
Newt Gingrich
I’ve said that even though the election was painful for Republicans, there was something comforting about the outcome during America’s 250th anniversary.
All that money, all those efforts by billionaires and political consultants, and voters simply rejected them.
In Texas, Republicans had enormous financial and advertising advantages, yet voters went in a different direction.
It raises serious questions about the political consulting profession and the assumption that enough money and advertising can produce a desired election result.
Your institute began tracking AI-generated election advertising in 2024. What did you learn?
Taylor Barkley
We established a fairly robust tracking system during the 2024 election cycle, looking for reported instances of AI-generated political advertising.
One of the most prominent examples involved a fake Joe Biden robocall in New Hampshire that provided voters with misleading voting information. The FCC ultimately imposed a substantial fine on the individual responsible.
That example illustrates something often missing from the debate: Existing laws already prohibit certain forms of election deception.
We also examined AI-generated videos and advertisements more broadly. Despite the warnings, we found no discernible impact on election outcomes.
Other researchers studying AI-generated election content internationally reached similar conclusions. There simply wasn’t evidence supporting the claim that artificial intelligence had taken control of democratic elections.
Newt Gingrich
One of the organizations you examined was the Wesleyan Media Project. What did its research tell you?
Taylor Barkley
The Wesleyan Media Project has assembled a substantial dataset on political advertising.
Its early figures indicated that Republicans accounted for roughly 90 percent of spending on ads containing AI-generated video. Yet that spending didn’t deliver the electoral results Republicans wanted.
The number of AI-generated advertisements has increased significantly, but people still vote according to their beliefs.
They rely on trusted networks, familiar sources, and their own judgment when evaluating political information.
Newt Gingrich
One of the great mistakes people make about politics is assuming voters are simply waiting for the right advertisement to tell them what to think.
There’s an underlying conversation happening all the time. People talk with family members, friends, and colleagues. They develop their own sense of what they want and what they don’t.
I’ve been studying Ford’s development of the Edsel. The company invested heavily in research and focus groups, yet the public rejected the car. Then Ford introduced the Mustang, which became an extraordinary success.
It’s an example of the public making its own judgment, regardless of what the experts predicted.
I think something similar happened during the midterms. Voters were responding to the cost of living, the war in Iran, and dissatisfaction with Washington. No amount of advertising could necessarily change those underlying concerns.
Taylor Barkley
That’s an important distinction between expert opinion and public judgment.
The argument that sophisticated technology can automatically manipulate the public assumes people don’t evaluate what they see.
But they do.
People have their own experiences, beliefs, and trusted relationships. They also have different priorities from the people producing political advertisements.
Even a highly convincing AI-generated video must reach the right audience, at the right time, through a credible source, and persuade people to take action.
That’s much harder than simply creating the video.
Newt Gingrich
I think it’s important to distinguish between different uses of artificial intelligence.
AI can improve the visual quality of an advertisement, create backgrounds, or enhance existing material. That’s essentially an extension of editing technology.
But that’s very different from creating a video of a political candidate saying something that person never said.
One is a creative tool. The other is an attempt to construct an alternative reality.
Taylor Barkley
Exactly. The law should focus on the deceptive conduct or illegal content, not simply the technology used to produce it.
AI can be a tremendous advantage for smaller campaigns that cannot afford expensive media consultants.
About 30 states have laws governing AI use in elections. Some require labels identifying AI-generated content.
But research suggests those labels often don’t accomplish what policymakers expect. People may ignore them, and in some cases, they can obscure the message itself.
We shouldn’t assume every use of AI in political communication is harmful. There are legitimate, valuable uses for this technology.
Newt Gingrich
Let’s talk about what happened in Maine with Troy Jackson. What does that example tell us about AI-generated political advertising?
Taylor Barkley
An advertisement appeared shortly before the election that portrayed Jackson doing something he had not actually done.
It had many of the characteristics that critics warn about: deceptive content, realistic imagery, and timing designed to influence voters before they could fully evaluate it.
Yet it didn’t produce the intended result.
In fact, the advertisement appears to have backfired.
That’s an important lesson. Even when someone uses AI in precisely the way critics fear, the public doesn’t necessarily respond as the creators expect.
Newt Gingrich
And it may have made Jackson a more sympathetic figure.
We saw a similar dynamic in Texas, where attacks involving AI-generated content didn’t produce the expected outcome.
It should cause political professionals to rethink some of their assumptions about voters.
Taylor Barkley
Absolutely. Candidates matter. Telling the truth matters. Being genuine matters.
There is evidence that voters are skeptical of politicians who rely heavily on deceptive AI-generated advertising.
Rather than being an unstoppable political weapon, these tactics can become a liability for the people using them.
Newt Gingrich
So you’re suggesting that, for most people, deepfakes simply don’t work?
Taylor Barkley
That’s what the available research and real-world examples suggest.
People form opinions through many different channels. They read, watch, listen, and talk with people they trust.
One area I’d like researchers to examine more closely is the role of podcasts and long-form audio in shaping political opinions.
When someone has listened to a podcast host for months or years, they’ve developed a sense of that person’s views and personality. It’s much harder for a last-minute AI-generated message to overcome that relationship.
Newt Gingrich
Although as AI becomes more sophisticated, it could probably imitate me and even create a hologram that looks like me.
Taylor Barkley
That’s true. You could create a fairly convincing artificial version of Newt Gingrich hosting a podcast.
But there would still be a real Newt Gingrich who could tell his audience that the imitation wasn’t genuine.
I think people want the authentic version.
It’s similar to the difference between mass-produced furniture and something handcrafted. Both have value, but authenticity and personal effort still matter.
As AI becomes more sophisticated, I believe our desire for genuine human connection will become even more important.
Newt Gingrich
Let’s turn to the First Amendment.
Courts have been examining laws that regulate AI-generated political content. What is the current legal situation?
Taylor Barkley
About 30 states have enacted laws governing artificial intelligence in elections, and some have faced constitutional challenges.
Courts have raised First Amendment concerns about restrictions that are overly broad or that compel political speakers to include certain disclosures.
Political satire and parody have a long history in America.
Think about political cartoons, exaggerated images, and humorous attacks on public officials. These have been part of our political culture for generations.
Artificial intelligence provides new ways to create that kind of expression, but that doesn’t mean the government should automatically prohibit it.
The danger is that fear of AI could lead policymakers to restrict legitimate political speech.
Newt Gingrich
So you would rather accept the risk of some deception than allow the government to micromanage political communication?
Taylor Barkley
I think the trade-off is too great when government begins deciding what information citizens should be allowed to communicate.
We have other ways to address deception.
People rely on family, community, trusted sources, and education. Public awareness of AI-generated content is also increasing.
There are risks associated with this technology, but there are tremendous benefits as well.
We should preserve an open environment for innovation while holding people accountable when they actually violate the law.
Newt Gingrich
America has a long tradition of tolerating political humor, exaggeration, and even harsh personal attacks.
During Grover Cleveland’s presidential campaign in 1884, his opponents used a mocking chant about his personal life. After Cleveland won, his supporters turned the chant into a celebration of his victory.
Political satire isn’t new. It’s been part of American democracy for a very long time.
Now, with Democrats controlling both the House and Senate, what do you expect Congress to do about artificial intelligence?
Taylor Barkley
I expect several areas of activity.
One will be proposals concerning liability for AI developers and the people deploying AI systems.
Our view is that people should be held accountable for illegal actions, regardless of whether they use artificial intelligence to carry them out.
If someone uses AI to steal money, the crime is stealing money. The technology is simply the tool.
I also expect congressional oversight hearings involving technology executives, greater attention to privacy regulations, and proposals concerning child safety.
There will likely be considerable debate over how much regulation is appropriate and how those rules could affect America’s ability to lead in artificial intelligence.
President Trump may be skeptical of proposals that interfere with American technological leadership.
Newt Gingrich
Let me ask one final question.
Like me, you’re an optimist about technology. Why?
Taylor Barkley
Because the history of technological progress is overwhelmingly a history of improving human lives.
Technology has reduced suffering, expanded educational opportunities, improved transportation, and created possibilities that previous generations couldn’t imagine.
My wife is alive because of advances in medical technology. That’s something I’m personally grateful for.
Artificial intelligence is an extraordinarily powerful technology, and I believe it will produce benefits beyond what we’ve seen from many previous innovations.
I’m the father of three young children, and I’m excited about the future they’re going to experience.
There are unknowns, and uncertainty can be frightening. But previous generations confronted technological change with optimism and ultimately benefited from it.
I believe we should approach artificial intelligence with that same spirit.
Newt Gingrich
I recently wrote a book called America’s Golden Age, and your perspective is very much in line with what I wanted to communicate.
I’ve become frustrated with the constant negativity surrounding technological change.
I used to tell people that if Thomas Edison invented the electric light today, the first headline would probably be about how the candle-making industry was threatened.
Throughout history, technology has empowered people, enriched societies, made us safer, and helped us live longer lives.
That’s something worth celebrating.
Taylor, thank you for joining me.
Taylor Barkley
Thank you so much for having me. This was a lot of fun.
About the Guest
Taylor Barkley is Director of Federal Government Affairs at the Abundance Institute, a nonprofit focused on creating a policy and cultural environment that supports emerging technologies, including artificial intelligence and energy innovation.
He previously held positions at the Center for Growth and Opportunity at Utah State University, Stand Together Trust, the Competitive Enterprise Institute, and the Mercatus Center at George Mason University.
His writing has appeared in Scientific American, USA Today, and IEEE Technology and Society Magazine. He also co-authored the Abundance Institute’s Tracking AI in the 2024 Election series.
Follow Taylor Barkley on X: @TaylorDBarkley.
You can manage your subscription preferences to choose the updates, newsletters, and alerts you want to receive on the website.





