This April, U.S. federal prosecutors charged an Army sergeant with using sensitive classified information to bet on Polymarket that U.S. forces would enter Venezuela and remove Maduro from power. About the same time, Kalshi disclosed that it had fined and suspended three congressional candidates for five years after they traded on prediction markets tied to their own candidacies.
The following month, CBS’s 60 Minutes broadcast a report examining suspected insider trading and other irregularities in prediction markets, including nine connected anonymous Polymarket accounts that made 80 bets on U.S. military operations, with a remarkable 98 percent win rate. And in July, Spotify reported removing more than half a million artificial streams that had distorted its U.S. chart, but only after Kalshi had already settled a $3 million prediction market event using the uncorrected results.
The series of reported market irregularities and controversies risks undermining trust in the fairness and legitimacy of prediction markets. They also appear more immediately consequential than the ongoing debate over how prediction markets should be classified, including whether they constitute gambling, how they should be regulated under U.S. law, and by whom.
More fundamentally, these and other events expose a growing gap between what prediction market providers say their products do and the social benefits they claim, and how these markets actually operate and the harms they can create. Prediction market providers and supporters make implicit and explicit claims about market integrity, fairness, “collective intelligence,” and truth. But product architecture and operational realities are producing behaviors and outcomes that undermine those claims.
Representation Gaps & Trust
We can call these gaps between product claims and operational realities “representation gaps”. When they proliferate and grow large enough, they undermine public trust and can constrain future growth and room to innovate. Think of representation gaps as producing a kind of BS alarm. When it rings loud enough, intense public outrage, political pushback, and regulatory constraints tend to follow.
In prediction markets, these representation gaps are unusually visible. Architecture choices around contract design, settlement rules, identity controls, incentives, and information asymmetries combine to produce observable real-world consequences. That makes it possible to compare what these markets claim, or are perceived by users, to deliver with actual outcomes.
Operators must mind the growing gaps between what users think they are betting on, what the contract actually says, the event that occurs, the evidence used to determine the outcome, and the market prices that users and outsiders may interpret as probabilities.
Those gaps are widening. Continuing with business as usual is itself becoming a risky bet.
Businesses routinely make implicit and explicit representations about their products and how they operate, including claims about purpose, integrity, transparency, safety, and control. Digital products operate through an overall and sometimes opaque product “architecture” that includes software code, rules, incentives, dispute resolution procedures, trust and safety measures, and other components and processes.
These elements work dynamically beneath the surface to influence how a product operates and the behavior it encourages. The architecture evolves over time in response to competing priorities, including growth, user experience, privacy, safety, compliance, and liability management. Public representations tend to emphasize one set of priorities. The underlying product architecture often reveals and favors others.
These gaps are rarely fully conscious or deliberately deceptive. More often, they emerge as the byproduct of competing internal objectives and accumulated product decisions, until someone examines them closely or a scandal brings them into public view. Privacy policies, for example, may promise user control while default settings and operating practices obscure data sharing and undermine meaningful choice. The details differ across digital products, but the pattern persists.
In prediction markets, outsiders can more readily detect gaps because trading records and, in some markets, blockchain activity reveal patterns such as account concentration, connected wallets, and long-shot contracts with anomalous win rates. That visibility invites greater public scrutiny. What’s more, a market price may reflect not only the probability of an underlying event, but also “resolution risk.” That reflects uncertainty about how specific contract language, evidence, and adjudication rules will ultimately determine settlement. Sophisticated traders may already be trading the rules and adjudication playbook, while more casual observers interpret the market price as the probability that the underlying event will occur.
Beyond Legal Compliance
Prediction markets don’t exist in a legal vacuum. Federal law already prohibits fraud and market manipulation in regulated prediction markets. States are also challenging claims of exclusive federal regulatory authority, arguing that some prediction markets fall within their traditional authority to regulate gambling.
In May 2026, Minnesota became the first state to enact an explicit ban, making the operation or facilitation of covered prediction markets a felony. The CFTC immediately sued to block the law, calling it the most aggressive state challenge yet to the federal regulatory regime. A federal judge temporarily blocked the law before it was scheduled to take effect on August 1. Other states have pursued their own legislation and enforcement actions. Nevada, for example, has asserted that certain sports, election, and entertainment event contracts require compliance with state gaming law and has obtained state-court injunctions against several operators. The jurisdictional fight may last years, but it doesn’t resolve the underlying concerns driving regulation.
Trading on material nonpublic information may violate federal law when the information is obtained or used through fraud or deception, including trading in breach of a pre-existing duty of trust or confidence. These prohibitions differ in important respects from, but also echo, traditional securities-law insider trading doctrine. It’s worth noting that the Commodity Exchange Act and CFTC rules, including Rule 180.1, do not impose a general duty to disclose material nonpublic information merely because a trader possesses it. Indeed, the CFTC has expressly recognized that derivatives markets have historically permitted trading on lawfully obtained nonpublic information in circumstances where other law doesn’t prohibit it.
But legal permissibility is not the same as market legitimacy and perceived fairness. Regardless of where legal boundaries settle, prediction markets will be judged on whether they are perceived as fair and trustworthy. Operators shouldn’t wait for others to define governance and operating boundaries. It’s in their own interest to more assertively police their markets and demonstrate the integrity and fairness they promise.
This is part 1 in a 2 part series. Stay tuned for the next installment.
