The fluorescent lights of a late-night campaign office have a specific, humming exhaustion. It is the sound of stale coffee, the crinkle of takeout bags, and the frantic tapping of keys. For decades, this was the heartbeat of American democracy—human beings burning the candle at both ends to win over a skeptical public. But in the quiet corners of Eric Swalwell’s world, a different kind of pulse is beating. It isn’t made of blood and adrenaline. It is made of silicon.
Representative Eric Swalwell, a man who has spent years in the high-stakes theater of Washington and is now eyeing the governorship of California, has a side gig. That phrase usually conjures images of an Uber driver or a weekend woodworker. For Swalwell, the reality is far more surreal. He is a co-founder of a company called dotComm. Its product? An artificial intelligence designed to mimic the nuances of political communication.
We are no longer talking about the future. We are talking about the present, where the line between a representative and a representation is beginning to blur.
The Ghost in the Machine
Consider a hypothetical voter named Maria. She lives in a cramped apartment in San Jose, juggling two jobs and a mounting pile of bills. She receives a text message from a candidate. The tone is empathetic. The policy points are specific to her neighborhood. The response to her skeptical follow-up is lightning-fast and eerily human. Maria feels heard. She feels like someone in power actually sat down to type those words to her.
Except, they didn't.
A machine analyzed Maria’s data, calculated the most effective linguistic emotional triggers, and generated a response in milliseconds. Swalwell isn't just using this technology; he is selling it. He is the pitchman for a world where the grueling labor of human connection is outsourced to an algorithm.
The core of the issue isn't efficiency. Efficiency is the buzzword used to mask the erosion of presence. When a politician uses an AI "twin" to handle the "drudgery" of constituent engagement, they are making a profound statement about what they value. They are saying that the act of listening is a bottleneck that needs to be widened by code.
The Ethics of the Avatar
Swalwell’s venture into the tech sector isn't happening in a vacuum. California is the forge where the world’s digital tools are hammered out. For a gubernatorial candidate to be both the regulator and the salesman of AI campaign tools creates a hall of mirrors. How can a leader pass laws to protect citizens from deepfakes or algorithmic bias when their own financial portfolio depends on the proliferation of those very systems?
It is a conflict of interest that feels uniquely twenty-first century. In the past, we worried about politicians being "bought" by oil companies or big banks. Now, we have to wonder if the politician themselves is becoming a proprietary software.
The technology behind dotComm seeks to capture the "voice" of a candidate. It learns the cadence, the favorite metaphors, and the specific policy stances. But a voice is more than just a pattern of words. It is the weight of a person’s lived experience. When you remove the person, you are left with a hollow shell that sounds like truth but lacks the accountability of a heartbeat.
The Invisible Stakes of Efficiency
Politics has always been a game of theater. We know the speeches are polished by committees. We know the tan is sometimes out of a bottle. Yet, there was always a tether to a physical reality. You could look a candidate in the eye at a town hall. You could catch them in a moment of unscripted frustration or genuine warmth.
If AI begins to handle the bulk of "engagement," that tether snaps.
There is a psychological cost to being "perfectly" represented by an AI. When the machine never makes a typo, never loses its temper, and always has the perfect statistic ready, it creates a standard that no human can meet. It turns the messy, beautiful, frustrating process of democracy into a streamlined consumer experience.
But democracy was never meant to be a product. It was meant to be a process.
Swalwell’s defenders argue that this is simply the evolution of the trade. They point to the printing press, the radio, and the television. Every leap in communication technology was met with fear, and every one was eventually absorbed into the status quo. They argue that if a candidate can reach ten times as many people through an AI, then more people are being served.
This logic holds up only if you believe that "being served" is the same as "being seen."
The Laboratory of California
California serves as the world's beta test. What happens here ripples across the globe. By embedding himself in the AI industry while seeking the highest office in the state, Swalwell is testing the boundaries of what the public will tolerate. He is betting that we are so tired, so overwhelmed by the digital noise, that we won't mind if the person shouting the loudest is actually a script.
He isn't alone. Other candidates are watching. They are waiting to see if there is a backlash or if the convenience of the tool outweighs the creepiness of the concept.
If the "Swalwell Model" succeeds, the 2026 election won't just be a battle of ideas. It will be a battle of optimization. The winner won't be the person with the best vision for the future, but the person with the most sophisticated data-scraping, response-generating engine.
The Human Residue
What is left for us? If the candidates are using AI to talk to us, and we eventually use AI to filter their messages, the entire political conversation becomes a dialogue between two machines. Humans become the bored spectators of our own governance.
We find ourselves at a crossroads where the most "effective" way to lead is to not be there at all.
There is a deep, quiet irony in a candidate for governor—a role that requires the ultimate human touch—spending his off-hours perfecting a way to automate that touch. It suggests a fundamental misunderstanding of what people are actually looking for in their leaders. We don't want a perfect, 24/7 response bot. We want someone who knows what it feels like to be tired, to be wrong, and to be real.
The digital doppelgänger might be able to quote every statistic in the book. It might be able to draft a policy paper in four seconds. It might even be able to win an election. But it will never be able to look a grieving mother in the eye or stand in the wreckage of a wildfire and feel the heat.
As the campaign trail moves further into the cloud, we have to ask ourselves: how much of our own humanity are we willing to trade for a faster reply?
The hum of the campaign office continues. Somewhere in the code, a version of Eric Swalwell is working. It never sleeps. It never eats. It never doubts. It is the perfect candidate, and that is exactly why it should make us very, very afraid.
