As vote tallies in the Los Angeles mayoral election trickled in slowly over the last week, unsubstantiated claims exploded on X that a fraudulent plot was underway to deprive the MAGA-backed former reality TV star Spencer Pratt the second-place slot to advance to the November runoff against incumbent Democrat Karen Bass.

A portion of these unfounded conspiracy theories pointed to changing betting odds for the three top candidates on prediction market sites Kalshi and Polymarket to suggest something sinister is afoot with the vote count. Some influencers supercharging such fraud claims online did so in posts sponsored by the companies themselves.

"They are actually doing it. They are counting votes until SPENCER LOSES. Someone DO SOMETHING," Trump-aligned influencer Mila Joy wrote to her half a million followers a day after the election as she reshared a Polymarket post with a graph showing that Pratt's betting odds were falling on the site.

"Is CA cheating to get Spencer Pratt out?" questioned commentator David Freeman, who posts under the handle Gunther Eagleman on X, as he shared a Kalshi post showing the odds between Pratt and progressive Democrat Nithya Raman. The Associated Press called the second-place spot for Raman on Monday afternoon after her vote share overtook Pratt's on Sunday.

At the bottom of both X posts, the words "paid partnership" appear in tiny font, a subtle reference to the millions of dollars Kalshi and Polymarket have pumped into programs that pay influencers to reshare corporate posts as a way to boost engagement.

The Los Angeles mayoral race is the clearest example yet of how prediction market posts about changing betting odds for candidates are being weaponized on X to sow doubt about the integrity of elections.

It's likely a preview of what's to come this year ahead of the midterm election. Kalshi and Polymarket are increasingly pervading ever more corners of daily life. Their rise has set off dozens of legal battles and raised novel questions about the ways betting on just about anything can have real-world consequences. Now it appears they are driving the latest battlefield in political misinformation wars on X.

"From the perspective of the influencer looking to get rich, their only job is to attract attention," Emerson Brooking, a disinformation expert at the Atlantic Council's Digital Forensic Research Lab, wrote in an email. "They will do this by sharing markets that align with what their audiences want to see. And if the betting markets are wrong, it is much wiser for them to allege fraud (and keep the lucrative promotions contract) rather than acknowledge that the gambling sites got it wrong."

In recent days, Kalshi and Polymarket have attempted to rein in some of their paid influencers. After NPR asked Kalshi about several partnership posts on Friday, the company said it told the influencers to take the posts down. Some of the posts, including Freeman's post questioning "CA cheating," have been deleted. Semafor first reported on Kalshi's crackdown.

On Monday, Polymarket told NPR it, too, is pulling back its sponsorship of some creators who were spreading election falsehoods. Joy's post is still live on X with the "paid partnership" tag, but the tag has been removed from posts by two other influencers paid by Polymarket.

"Companies shouldn't be paying people to spread misinformation," said Brendan Nyhan, a political scientist at Dartmouth College, who has reviewed the sponsored posts that flew across X. "In the Trump Republican Party, fraud allegations are going to be often received with a lot of enthusiasm, especially when people often get confused about the difference between the odds of someone winning and vote share."

Inside Kalshi and Polymarket paid partnerships 

Paying influencers as social media promoters is a type of "growth hacking" tech startups often deploy to maximize the reach of their brand in an attempt to drive more users to the services.

"It's a high-risk, high-reward situation," said Seton Hall University's Jess Rauchberg, who studies digital media culture. "But it's a strategy that gets people talking about the brand."

Kalshi and Polymarket have offered creators as much as $500 per post, according to two people who formerly worked on partnerships at Kalshi and Polymarket and who were not authorized to speak publicly about the programs.

Inside Kalshi, the approach has sparked debate over what responsibility the company has when creators promote its site by spreading misinformation and other harmful content across X, according to the former Kalshi employee.

A Kalshi spokesperson confirmed on Monday that the company now prohibits anyone in its affiliate program from questioning the integrity of an election or undermining a legal ruling or official determination about an election.

Previously, the company took a mostly hands-off approach to what its affiliate creators posted to boost one of Kalshi's markets, according to the former Kalshi employee.

Before the recent controversy, one of the only times Kalshi cut ties with a paid creator over a post promoting the company was when one of their contributors posted to X celebrating the death of conservative political activist Charlie Kirk, the former Kalshi employee said.

Similarly at Polymarket, affiliate posts were given wide latitude, as long as the person posting plugged the company's markets, according to the former Polymarket employee. And there appeared to be little vetting of creators, with Polymarket tapping former Rep. Matt Gaetz as one of its paid contributors. The U.S. House Ethics Committee found Gaetz paid an underage girl for sex.

On Monday, Polymarket said that while it does not have language specifically banning creators from posting election-related disinformation, any post denying the result of an election would violate its rules against spreading false and misleading information.

Polymarket told NPR posts from two of the creators it works with have lost the "paid partnership" tag. It has not asked creators to delete any posts, but told them about the company's content guidelines.

While the company would not specify which creators, NPR confirmed "paid partnership" tags have been removed from Jun. 4 posts by right-wing influencers Benny Johnson and Kangmin Lee sharing the same Polymarket post about Raman's rising odds on the betting site.

Seton Hall University's Rauchberg said the crackdowns are just the latest example of how the rival companies are constantly trying to one-up each other.

"They want to spread this rhetoric that 'Kalshi is for everyone, Polymarket is for everyone,'" she said. "They want to give the impression that they don't have a political affiliation, but consumers are becoming more savvy that both companies are engaging in a type of 'purity politics,' each trying to outdo the other over which is the best app to use."

Not disclosing whether a social media post was sponsored is illegal under rules the Federal Trade Commission adopted in 2024. The Trump administration has not rolled back these rules, but it has also not announced any enforcement actions.

Why California vote counting attracts fraud claims

The Los Angeles mayor race was particularly vulnerable to becoming the focus of election conspiracy theories for a number of reasons. Prediction market data may have been one of them.

Pratt, an outsider candidate who received outsized attention and engagement on X, was favored for second place on betting markets on both Kalshi and Polymarket's sites in the days before the election — even when the largest polls of likely voters showed him in third place.

While the city's mayor is a nonpartisan office, a registered Republican like Pratt faced a challenge in heavily Democratic Los Angeles. But some social media commentators cited his favorable betting odds as evidence he could reach the November runoff.

Posts about what betting markets are saying about a candidate can confuse voters who may not understand the difference between betting behavior and a poll, said Zarine Kharazian, a Ph.D candidate at the University of Washington's Center for an Informed Public, who studies online rumors related to elections.

"It runs the risk of confusing people into thinking that, 'Okay, these markets have the pulse on public sentiment about the election and who's going to win,' when that's not necessarily the case," Kharazian said.

Heading into the Jun. 2 primary, election experts were already worried that California's notoriously slow ballot count would provide the opportunity for baseless fraud allegations to blossom.

A large portion of voters in the state use mail-in ballots, a form of voting President Trump has tried to associate with fraud. Election officials must verify mailed-in and dropped off ballots, making them slower to count. The state accepts ballots that are postmarked on the day of the election that arrive within seven days.

Ballots that are counted later in the process typically skew Democratic since more voters from that party embrace voting by mail. This phenomenon has been the basis for unfounded allegations of fraud in recent years, including by Trump.

The challenge has been particularly stark this year because so many Californians waited until Election Day to drop off their ballots, said David Becker, executive director of the Center for Election Innovation and Research.

State officials are "doing what they've always done — counting all the ballots, under transparent observation from the candidates and parties, and reporting each batch as soon as they can," Becker said, "yet the profiteers and grifters are loudly echoing our foreign adversaries in spreading lies designed to delegitimize our transparent election process."

President Trump himself has claimed without evidence that there was fraud in the Los Angeles mayoral's race. He called the election race "rigged" in a Truth Social post early Monday, and wrote it was "not possible" for Pratt to lose to Raman after his initial lead when vote counting began. The first assistant U.S. Attorney for the Los Angeles area, Bill Essayli, announced on X days earlier that his office had multiple election fraud investigations underway.

Over the weekend, Essayli debunked one popular conspiracy theory circulating on X — that Pratt had received zero votes in a ballot count update — as false.

Election experts say the baseless fraud allegations in California do not bode well for the upcoming November midterm season.

"I think we're going to get punched in the face so badly on election denialism in November," said Stephen Richer, the former Republican recorder for Maricopa County, Arizona, who dealt with baseless fraud allegations in the aftermath of the 2020 election. He is now a legal fellow at the Cato Institute and a senior fellow at Harvard University's Kennedy School of Government.

Richer recalled that during the 2020 election, people trying to undermine the election results latched on to graphs that showed a blue line representing former President Joe Biden's totals suddenly jump higher as ballots were counted.

"And so now it seems that they're using these prediction market graphs to tell a similar story," Richer said.