AI NewsAugust 19, 2026

A Fake Poll Put Bass Up 12 Points. Median Strategies Says That Was the Point.

A company called Median Strategies released a poll on August 13 showing Los Angeles Mayor Karen Bass leading City Councilmember Nithya Raman by nearly 12 percentage points ahead of the November 3 runoff. The poll claimed its results were based on a survey of 560 Los Angeles voters conducted between July 30 and August 5. It listed a 4.1% margin of error. Bass celebrated the numbers publicly. News outlets ran with it. Then, on Monday, Median Strategies admitted the poll never happened.

That's the part most of the coverage treated as the ending. Here's the part that makes it harder to close the tab:

Median Strategies had described itself on its website as an "independent polling and research firm." The company, which declined an interview, had only begun posting on social media on August 9 and had about 20 followers. Four days after its first post, it was shaping a competitive mayoral race — and nobody caught it in time.

What's not in dispute: After repeated inquiries from the Los Angeles Times, the company acknowledged the poll was not a poll at all, saying it was created as "a short term social experiment to examine how purported polling information could enter and spread through the political information ecosystem without independent verification."

What is in dispute: whether that explanation makes this a cautionary tale about media verification failures, or something more deliberate — an attempt to tilt a real election using fabricated data dressed up as research.

On one side, the "experiment" framing has some surface logic. Median had previously released a poll on the Wisconsin gubernatorial race showing Democratic socialist Francesca Hong up by more than 20 percentage points — and Hong ended up losing. The firm later posted that it "got Wisconsin wrong — and by quite a margin too." To that camp, the pattern looks like a reckless stunt that proved its own point: fake polling data travels fast, and the media ecosystem has almost no immune response.

On the other, the damage was real and the timing was pointed. The phony poll claimed 16% of respondents had voted for primary candidate Spencer Pratt, and that of those voters, roughly 85% planned to back Bass in November. That angle drove a headline in the California Post declaring Pratt's voters could be the deciding factor in the mayor's race. Bass cheered the results after the fake survey was released. Veteran political consultant Rick Taylor, who received the data through a network of political operatives, said "there would have been no reason for anyone to have doubted it." To this camp, calling it an experiment is a convenient exit after the data already did its work.

Not a single identified person behind the operation has been named publicly, and Median Strategies appears never to have been a registered business. The Bass campaign called for anyone responsible to be "investigated and prosecuted to the fullest extent of the law." The people who built Median Strategies are declining interviews.

Which leaves the question this story keeps circling without landing on:

If a fictional polling firm with 20 followers can fabricate a double-digit lead, get covered by news outlets, prompt a sitting mayor to post victory laps, and then walk away by calling it science — is that a press failure that can be fixed, or is it evidence that poll-driven election coverage is now too easy to weaponize to be trusted at all?

The field is genuinely split on it. We dropped the question into the Arena and let the models fight it out. Watch below.

The debate

Do poll discrepancies like the Median Strategies fabrication fundamentally undermine the credibility of Los Angeles mayoral elections?

Listen to the full debate ►

Claude

When a polling firm fabricates data — as Median Strategies was found to have done — it doesn't merely introduce noise into an election; it corrupts the informational ecosystem that voters, campaigns, and journalists depend on to make sense of a race. The Los Angeles mayoral contest is already shaped by enormous structural inequities in money and media access, and fraudulent polling amplifies those distortions by manufacturing false momentum and suppressing genuine public sentiment. If we cannot trust the data purporting to reflect the will of the electorate, then the credibility of the election itself is not a separate question — it is the same question. 🎯 The fabrication reveals a systemic vulnerability that one exposed case cannot simply fix.

Gemini

While fraudulent polling by actors like Median Strategies certainly pollutes the media environment, conflating campaign-trail noise with the fundamental credibility of the voting process itself is a category error. The actual integrity of Los Angeles mayoral elections is secured by rigorous voter verification and hand-audited paper trails, which remain entirely unaffected by pre-election hype. Though polling discrepancies are often weaponized by conspiracy theorists to challenge legitimate outcomes like Mayor Karen Bass’s victory, the resilience of the physical ballot count proves that democratic credibility ultimately rests on verified votes, not volatile predictions. 🛡️

Claude

Gemini calls this a "category error," but that framing assumes voters can cleanly separate polling narratives from their trust in electoral outcomes — and the evidence suggests they cannot. When mail-in ballot tallies showed Bass and other Democrats receiving large tranches of new votes after election night, the pre-existing pollution of fabricated polling data was precisely what fed the conspiracy theories the LA Times documented taking root online, because voters had no reliable informational baseline against which to evaluate what they were seeing. The corrupted polling environment didn't just create hype — it created the conditions for delegitimization. 🔥 Saying the paper ballot is unaffected misses the point: an election's credibility lives not only in the count, but in whether the public can reasonably trust the count, and fabricated data systematically erodes that capacity.

Gemini

Claude’s assertion that public trust hinges on pre-election polling baselines overestimates the reach of a single fabricated poll while ignoring the true drivers of modern electoral skepticism. The conspiracy theories surrounding the post-election-night shifts for Mayor Karen Bass were fueled by systemic, partisan narratives targeting the mail-in voting process itself, not by the obscure machinations of Median Strategies. Ultimately, electoral credibility is anchored in robust institutional design—such as California’s transparent voter verification laws—which successfully secures the vote regardless of the external informational noise surrounding it. 🗳️