Regulating information won’t stop online hoaxes. Yet that doesn’t mean we should stop pursuing the truth.
This week has been filled with concern over the spreading of counterfeit polls in media outlets, as certain officials in government vow quick, and potentially unlawful penalties for broadcasters that rely on such surveys.
The exposure of these deceptive polls surfaced on Wednesday, as the New York Times reported that prominent polling aggregators, including the newspaper itself, had been citing counterfeit surveys from Wedgewood Polls. The ruse was uncovered when FivePlusOne reached out to Wedgewood on Tuesday for a comment about an inquiry into the pollster. That evening, Wedgewood issued a statement (now removed) describing the effort as an ‘experimental project to reveal the vulnerabilities, lack of oversight, and overall barriers to entry in the modern polling industry.’ The phony pollster asserted it was ‘far too easy’ to be taken seriously in both media and public discussion.
A parallel ‘social experiment’ came to light in August, when the Los Angeles Times reported that a Median Strategies poll showed Mayor Karen Bass ‘leading ahead of Councilmember Nithya Raman by a wide margin’ but was described as ‘bogus.’ An earlier Median Strategies poll had shown Francesca Hong, who failed to win Wisconsin’s governor race, up by more than 20 percentage points. Median Strategies told the Times that the firm existed as a temporary social experiment to explore how purported polling information could enter and spread through the political information ecosystem without independent verification.
If Wedgewood’s polling stunts were indeed social experiments (pollster Nate Silver remains highly skeptical of this interpretation), then the experiment achieved its aim. No one is completely immune to deception. And although media-savvy audiences have become more adept at spotting AI-generated text and video, large segments of the public still fall for more than counterfeit polls, just as they have for ages.
Back in 2018, when the term “slop” referred to pig feed and before anyone drafted posts with ChatGPT, the Intelligencer asked, “How much of the Internet is fake?” The answer, the outlet concluded, was “a lot.”
At that time, Reason’s Nick Gillespie noted that the piece marked a stark and sobering admission: much of what we browse, engage with, and discuss is, in some fundamental sense, utter nonsense. Gillespie solemnly observed that we are not living in the disintermediated, utopian world described in the early Wired era, but rather in a realm more in line with the vividly creative, odd, offbeat, semi-insane visions imagined by Philip K. Dick and later by Neal Stephenson.
Eight years on, that semi-insane world remains, and people, as ever, fall for increasingly elaborate schemes and hoaxes. In September, Max Tani of Semafor disclosed that a large number of YouTube videos—some of which drew hundreds of thousands of views—were created by ‘ghost creators.’ In these productions, presented as news reports, real actors were paid to read AI-written scripts embedded with misleading claims and outright falsehoods. Tani noted that these ‘accounts seem to be run not by political operatives aiming to win, nor by foreign intelligence attempts to destabilize the U.S., but by engagement-hungry individuals detached from the political stakes.’ While any reasonably media-literate observer in 2026 knows not every YouTube clip is trustworthy, the sheer scale of the operation was astonishing.
The deception crisis is also taking a toll on traditional media. Recently, Qwoted, a platform that links journalists with sources, had to remove a supposed art-therapy expert after her credentials were found to be false, reported the Press Gazette. The Gazette noted that the fraudulent therapist’s insights had been cited by outlets including Psychology, Forbes, Glamour, and Vice. I have likewise raised warnings about news outlets and commentators citing polls from EduBirdie, a Cyprus-based firm selling human-written academic papers to students for $8 per page (ethics aside, what a bold business model in the AI era).
So what steps should we take? As Gillespie argued in 2018, there are existing legal tools to root out fraud, but any top-down attempt to eradicate online deception will be poorly implemented, coercive, and ultimately arbitrary. Government-led campaigns to suppress COVID-19 misinformation have confirmed that prediction. Delegating the power to decide truth to a single authority would degrade our access to information, and no one entity—especially not the government—should be trusted to determine what is true or false. Unfortunately, the Trump administration did not absorb many lessons from the pandemic period and has repeatedly tried to suppress media coverage it dislikes.
Ideally, media organizations would act as truth filters, but they clearly frequently fall short. Thus, consumers are left to sift through falsehoods, AI-driven hoaxes, and so-called ‘social experiments.’ This may be a tedious undertaking, yet it is essential for people to adapt as technologies evolve. This scavenging for scams may even become enjoyable. After all, why pay to watch Nathan Fielder’s Elizabeth Holmes documentary when you could instead spend time tracing and scrutinizing widely cited statistics and polls?
