The AI Authoritarian Threat

Your weekly excerpt from one of my books. This week: "The Last American President: A Broken Man, a Corrupt Party, and a World on the Brink"

The Last American President: Over the next few months the entire book will be here, new chapters posted every Sunday, for subscribers to read at no cost. If you want to get a physical book to mark up or share with others, just click on the picture above, visit your local bookstore, or check your favorite online seller.

The AI Authoritarian Threat

Pope Leo XIV labeled AI one of the main threats facing humanity, saying it poses challenges to human dignity, justice, and labor.21 He’s right, but it’s even worse than that; AI represents, unless it’s rigorously regulated, a threat to democracy itself.

In every generation, the enemies of democracy change costumes, nations, and languages, but their playbook remains eerily familiar. They lie, divide, intimidate, and exploit every available tool to consolidate power. In the 1930s they used newspapers and radio, in the 2010s it was social media, and now, in 2025, the newest and most dangerous weapon in the authoritarian arsenal is artificial intelligence.

Make no mistake: AI isn’t just another “new technology.” It’s power, scaled. And in the hands of the hard right that is trying to end our republican form of government, it has the potential to become the most effective tool for dismantling democracy ever invented.22

Authoritarians—whether MAGA-aligned in the United States or part of the global movement that includes Putin, Orbán, Modi, MBS, and others—­are not blind to the potential of AI. They understand it instinctively: its ability to pretend to be human, to deceive, to surveil, and to dominate. While progressives and democratic institutions are scrambling to get a handle on its implications, authoritarians in America, Russia, and around the world have already started weaponizing it with devastating efficiency.

A single AI machine can now generate millions of personalized political messages in seconds, each calibrated to manipulate voters’ specific fears or biases. It can (and currently is being used to) create entire fake news outlets, populate them with AI-generated journalists, and flood your social feed or web search with content that looks real, sounds real, and feels familiar, all without a single human behind it. Imagine the power of Joseph Goebbels’s propaganda machine, but with superintelligence behind the wheel and zero friction. That’s where we’re heading in the 2028 presidential election.

And that’s just the beginning.

Authoritarian regimes can—and already are—using AI to surveil and intimidate their citizens. What China has perfected with facial recognition, social media, and loyalty scoring, MAGA-aligned figures in the United States are rushing to adopt and adapt. Right-wing sheriffs and local governments could soon use AI to track protesters, compile digital dossiers, and “predict” criminal behavior in communities deemed politically undesirable. If the government knows not just where you are, but what you’re thinking, organizing, or reading—­and it can fabricate “evidence” to match—freedom of thought (much less freedom of expression) becomes a quaint memory.

This isn’t theoretical. In 2024, Republicans deployed AI-generated robocalls impersonating Joe Biden telling voters to stay home, and millions did. In the next cycle, it’s safe to predict that we’ll see entire portions of election campaigns waged by AI bots masquerading as voters, influencers, news media, and even public officials.

The goal here for the hard right that doesn’t embrace democracy but wants America to become an authoritarian state isn’t just to win; it’s to delegitimize the democratic process itself as Orbán and Putin have done. Because once trust is broken—once people believe that “both sides lie” or that “you can’t believe anything anymore”—then, inevitably (history tells us), strongmen step into the void with promises of order, purity, and salvation.

And when they do, AI will also be there to help them enforce the new “order” in their “orderly society” where dissent has become a crime and fear of speaking out stalks the land.

Imagine a future where police departments outsource their decision-­making to “neutral” algorithms, algorithms coded with the biases of their creators, like Musk is doing by training his Grok AI on X. Where AI-driven court systems deny permits, benefits, or even due process based on Majority Report–style “behavioral profiles.” Where (like in China today) loyalty to the regime is rewarded with access, and dissent is flagged by invisible systems you can’t appeal or even know about for sure.

That’s not democracy. That’s techno-feudalism, wrapped in a red-white-and-blue flag.

If we allow the billionaires who fund the hard right to continue merging political power with unregulated AI, we will see the rise of a system where freedom is algorithmically rationed.

Elections will still happen, of course, but outcomes will be predetermined without our even realizing it. Dissent will still exist, but only in controlled pockets, and it’ll be easy to monitor and suppress through harassment, intimidation, and arrest. History books will be written, edited, and distributed by AI code optimized to tell the oligarch’s story while suppressing the true stories of America (particularly those of racial and gender minorities). In this Brave New America, the “news” will be whatever the regime’s AI decides you should see.

This is not science fiction. It is the logical endpoint of unregulated, authoritarian-aligned artificial intelligence in the hands of unaccountable billionaires and despotic governments.

So what can we do?

To start, we must treat the regulation of AI and the people who own/use/deploy it as a democratic survival issue. That means

· Banning the use of deepfakes in political ads

· Enforcing transparency on algorithmic decision-making

· Creating public, open-source alternatives to corporate-controlled models

· Creating disinformation-catching infrastructure as we would biological or nuclear weapons (that are also not just dangerous, but potentially civilization-ending)

· Demanding that social media outlets publish their algorithms so we can see how we’re being manipulated

And we must do it now.

Because history teaches us that once authoritarianism takes root, it rarely gives up power voluntarily. Particularly when (like with the innovations of the nineteenth and twentieth centuries) it has new tools that are more powerful than those who would protest and try to defend democracy. The longer we wait, the more embedded, autonomous, and intelligent these systems will become, and the more wealth and power their owners will accumulate. We’re not just fighting bad actors anymore; we’re fighting machines trained by them to think and behave like them.

The battle for democracy in the age of AI won’t be won with slogans or optimism alone. It will take law, oversight, courage, and above all, vigilance. I end every radio show with “democracy is not a spectator sport.” If we want to preserve the sacred right of self-governance we inherited from previous generations, we have no choice but to immediately recognize the existential threat in front of us and act with appropriate urgency.

This time, the fight isn’t just against the usual suspects.

This time, the algorithm is watching.