For as long as anyone can remember, pundits and pollsters have furiously weighed in with analysis and predictions on upcoming presidential elections. Of course, even though some will occasionally hit the mark, projecting election outcomes is the product of imperfect humans with their inherent flaws, subjectivity, and biases. So, as we dive headfirst into the age of artificial intelligence (AI), it is time to ask whether robots might be more capable than people of accurately forecasting the outcome of a presidential race encompassing a vast array of issues and the whims of 150 million voters in 50 states and DC. After all, if it can create breakthroughs in everything from the performance of household chores to fast-tracking new medicines to war strategy, AI should certainly be able to look into its crystal ball and tell us who will follow Donald Trump into the White House, right?
Going Where No Human Has Gone Before
While the outcome of the 2026 midterms and significant events to come in the 31 months before the 2028 presidential election are obviously unpredictable, AI has a unique capacity to sort almost endless amounts of information into a coherent prognostication. That is not to say every AI model will necessarily reach the same conclusion, as we will explore. But like economists whose predictions are wrong at least as often as they are right, we still find their analysis useful, so let’s dive into what various iterations of AI are predicting for 2028.
But first, let’s remember that AI models can also be biased based on how they are constructed by humans. For example, one iteration of AI might refer to Jan. 6 as an insurrection, another as a riot, and yet another as a disturbance, based on how they are programmed. Since it is impossible for us to understand the exact algorithms driving any built-in biases that might exist, let’s employ the same basic method as Liberty Nation’s Public Square and seek a consensus among some of the most prominent and widely accessed AI platforms.
Google’s Gemini cites political trends, strength of competition, exhaustive data from Polymarket, “the world’s largest prediction market,” among other factors, and concludes that Republican JD Vance will narrowly defeat Democrat Gavin Newsom 278-260 in the electoral college. Its summary logic: [Vance] is the betting favorite, Trump's heir-apparent (Vance picked as VP in 2024), and leads early GOP polls. "Incumbent-party fatigue might hit the Trump admin by 2028, but GOP base loyalty + Rust Belt appeal (from 2024 win) gives edge," Gemini notes.
Elon Musk’s Grok concludes that “Vance is the safest bet for the Republican side by far. The Democratic nomination is a true scramble, but Newsom currently looks like the slight favorite. The general election is a coin-flip right now.” Some other platforms have a different view of the eventual GOP nominee. On Claude AI, Marco Rubio is currently projected to win the GOP nomination and capture the 2028 election with 291 electoral votes to Gavin Newsom's 247, on the strength of “tight wins in states like Georgia and Pennsylvania.”
Perplexity, which draws from multiple AI models, concludes: “My best prediction, based on current market signals, is: Democratic nominee: Gavin Newsom, Republican nominee: JD Vance. Winner: Democrat, by a narrow electoral-college margin. In the current market view, the likeliest individual winner is JD Vance, but the party-wide favorite is Democrats.” That last sentence is confusing, but seems to imply that Democrats will be favored generically, but Vance will be the most popular personally.
AI Is Not God
All of these AI platforms emphasize that their level of confidence in their current prognostications is low, which is unsurprising for a race more than two and a half years away. As Grok states, “it will be decided by events between now and 2028 far more than today's polls. History shows these early pictures are often wrong (e.g., 2016, 2020).”
Robots in this instance are being used to filter information and use logic to reach conclusions, not to think in unconventional ways, e.g., would AI have predicted a Trump victory in 2016? Therefore, much common AI thinking about 2028, particularly in the forecasts of a very close election, is to be expected. But the fact that these predictions are unanimous about the identity of the eventual Democratic nominee does hold a certain significance. Newsom’s movie-star looks belie his well-documented deficiencies, so this collective belief about the California governor winning the nomination could well be interpreted as an indication of the overall weakness of the expected Democratic field. At the same time, few will be surprised that AI concurs with the general conclusion that JD Vance remains a favorite over Marco Rubio, but by a dwindling margin, on the Republican side.
Take these forecasts for what they are worth, which is not nothing. But at the very least, enjoy the great fun and infinite predictive capabilities of the space-age technology that will drive the future of the nation and the world.
.jpg&w=1920&q=75)
