I enjoy watching the Super Bowl as much for the commercials as I do the game. And I know I’m not alone.
Sure, the game might deliver more excitement in the moment. But the ads tell you something about where business and culture are heading.
After all, a 30-second ad at this year’s Super Bowl cost between $8 and $10 million. You don’t spend that kind of money unless you’re trying to cement your place in the mainstream and signal that you’re building something big.
This year, that meant a wave of AI ads. Startups and tech giants alike were pitching automation, copilots and digital assistants as the new norm. The AI theme was unmistakable, even though not every spot hit the mark.
But do you know what I didn’t see during this year’s Super Bowl?
There wasn’t a single ad for prediction markets.
And that wasn’t an accident. The NFL banned advertising from platforms like Kalshi and Polymarket for the entire 2025 season, despite these platforms growing fast and attracting billions in funding and mainstream attention.
In fact, the NFL specifically kept these platforms out of the Super Bowl broadcast, putting them in the same prohibited category as tobacco and firearms.
The league says it’s concerned with integrity. League officials argue that these markets lack the safeguards of regulated sports betting, including protections against manipulation and strict data rules.
And maybe for good reason.
Because prediction markets — especially when combined with AI — are becoming something much more powerful than a novelty bet on the future.
They’re becoming engines of collective intelligence.
AI Joins the Crowd
Prediction markets work on a simple idea.
Instead of asking experts to guess what happens next, you let thousands of participants trade contracts tied to outcomes. Prices move based on conviction and money on the line. Over time, the market aggregates information, incentives and sentiment.
This isn’t just a theoretical approach.
One analysis of Polymarket data found the platform was about 90% accurate in forecasting outcomes a month ahead of events and up to 94% accurate shortly before they occurred.
And we saw that dynamic play out in 2024.
During the presidential election cycle, more than $3.3 billion flowed through Polymarket contracts tied to the race, with industry estimates putting total market activity closer to $3.7 billion.
And as all that money moved, the market odds started to differ from what polls were showing.
As the election got closer, markets priced Donald Trump’s chances as much higher than Harris. Yet, many surveys at the time framed the race as essentially even.
Large traders leaned into those signals. One participant alone placed positions with potential payouts near $46 million, as probabilities shifted toward roughly 62% versus 38%.
Down the ballot, the same thing was happening. Candidates favored by market pricing went on to win about 89% of competitive Senate races.
Researchers studying the election noted how probabilities evolved in real time across months of trading activity, highlighting a responsiveness traditional polling structures struggle to match.
But when you add artificial intelligence into the mix, the dynamic evolves even further.
Researchers studying conversational AI-assisted forecasting found that groups collaborating through AI mediation predicted Major League Baseball outcomes with 78% accuracy, beating Vegas betting markets that landed at 57%.
Source: unanimous.ai
Again, this advantage didn’t come from AI predicting alone. It came from using AI to structure debate and sharpen human judgment.
And we’re seeing similar results elsewhere.
One study showed that when human forecasters had access to advanced language model assistants, their prediction accuracy improved between 24% and 28%.
Yet, fully automated models trying to predict financial markets still struggle. Many approaches barely break past the mid-50% accuracy range, and even advanced hybrid systems only push accuracy toward about 60%.
The pattern here is pretty clear.
AI isn’t perfect at forecasting, and neither are we. Machines miss context, while humans bring their own biases.
But when you put them together, accuracy improves. And that’s starting to have real-world consequences.
Prediction markets take a wide range of opinions and turn them into prices that reflect probability. AI then digs into that data, finds structure and highlights signals that people wouldn’t see on their own.
Once those signals exist, they can influence decisions across investing, operations, risk management and many other areas.
So I can see why the NFL is uneasy about polymarkets. These platforms don’t just surface information quickly. They reflect public sentiment in real time, and that can shape behavior.
For a league built on competitive integrity, that’s a risk it can’t afford to ignore.
Here’s My Take
Prediction markets proved that crowds can outperform experts. AI is now proving that when these crowds are structured and sharpened by machines, they can do even better.
And it’s hard to ignore where this is heading.
Intelligence is changing. We’re moving toward a world where people and machines think alongside each other in real time.
Forecasting is simply the first place we see it happening today. But I don’t believe this hybrid intelligence will stay confined to prediction markets.
Wherever important decisions depend on probabilities and incentives, combining human networks with machine intelligence could improve the outcome. I’m talking about things like capital allocation, supply chains, political strategy and corporate planning.
Which means prediction markets could eventually evolve into infrastructure for decision-making itself.
And when that happens, we might look back at today’s polymarkets debates the same way we look at early arguments about online trading.
The moment before adoption became a foregone conclusion.
Regards,
Ian KingChief Strategist, Banyan Hill Publishing
Editor’s Note: We’d love to hear from you!
If you want to share your thoughts or suggestions about the Daily Disruptor, or if there are any specific topics you’d like us to cover, just send an email to [email protected].
Don’t worry, we won’t reveal your full name in the event we publish a response. So feel free to comment away!
