In my previous post, I promised to share how our AI model was able to predict election outcomes with 100% accuracy, down to the percentage levels in swing states.
Let me explain how. It’s not only about AI; it’s about who is training and using the AI. In our case, as neutral observers, we already had the advantage of near-zero confirmation bias and a comfort with divergent data that almost no pollster or media house was willing to consider. I won’t go into the technical details because that’s our main ‘bread and butter,’ but I would like to share a simple theoretical framework I built on how to compute this: the ‘Common Sense’ Framework.
The goal of any electoral campaign is to build a ‘common sense’ around the candidate that the public can easily resonate with. The closer a candidate aligns with ‘common sense,’ the larger the landslide.
What does it take to build a common sense?
An extraordinary understanding of people
The ability to use data to visualize and demonstrate common sense to the public
Simplified messaging that is digestible
The right amount of intensity to present the “other” as defying common sense
Repeat. Repeat. Repeat.
Once you can quantify and compute this framework into your AI model, the results start to reflect the truest public pulse on elections and voting behaviour.
Elections—or, frankly, any narrative game—is really about winning over public common sense. A good leader aligns with the people’s common sense; a great leader generates ‘common sense’ in the people; and the greatest leader becomes ‘Common Sense.’
In the context of the U.S. elections, you’ll see just how distant Kamala’s campaign and narratives were from the public’s ‘common sense.’ The emphasis on gender change and transgender debates, along with continued involvement in Ukraine and Gaza as the U.S. struggled, defied the public’s ‘common sense.’ Trump, on the other hand, focused on immigration, taxes, the U.S. economy, and ending wars—all of which appealed to the common sense of not only his base but also a specific set of voters in swing states.
At the end of the day, it really comes down to humans, but AI can help us understand the public and its ‘common-sense’ in fascinating ways and, in doing so, help predict voting behaviour and much more.
I have a some experience running AI predictive models and even I thought your predictions off and wondered if I would hold them back because they appeared to be skewed but after seeing Trump sweeping all those states, it was a reminder that AI can help us check our own biases and sometimes that means getting a reality check.