Biased media coverage and the lab leak theory

Gain-of-function fearmongering and blame gets the glitzy NYT visual treatment as Republicans weaponize the lab leak theory again in attacks on Dr. Anthony Fauci

The lab leak theory is dominating the news cycle again following a contested Senate hearing and subsequent media discussion over the origins of SARS-CoV-2.

In the summer of 2024, the retired public health servant Anthony Fauci was dragged in front of Congress in a witch hunt to blame the origin of the pandemic on him. On the day, the New York Times published an op-ed of Alina Chan, a well-known activist on the topic with a vested interest through her book Viral and a non-profit she founded, with incredibly elaborate visual support from the New York Times visual and graphic artists teams.

Now, in the summer of 2026, the retired public health servant Anthony Fauci was dragged in front of Congress in a witch hunt to blame the origin of the pandemic on him. Soon after, the New York Times published an op-ed of Alina Chan, a well-known activist on the topic with a vested interest through her book Viral and a non-profit she founded, with incredibly elaborate visual support from the New York Times visual and graphic artists teams.

They say history might not repeat itself exactly, but it certainly rhymes. If we are indeed living in a simulation, somebody clearly messed up the copy paste. Unfortunately, our reality seems no less scripted sometimes. Anxieties and gain-of-function fearmongering of the past years are once again culminating in a recent - and very visually appealing, not gonna lie - op-ed by Dr. Alina Chan. It appears the lab leak influencer has once again been asked by the Grey Lady for another spin around the lab leak ballroom.

Since 2020, the lab leak idea in its many iterations of inconsistent allegations and speculations has become a staple of public and elite imagination, despite a complete absence of any recognizable evidence justifying such strong belief. Not to mention a large and growing scientific body of work and consensus strongly in contradiction with the idea of a research-related origin.

Just typing yet again this contextualizing paragraph about the current status of the “origin controversy” makes me cringe. Nothing has changed on this assessment in years. It is literally old news. Let me state it plainly:

The weight of scientific evidence for a zoonotic origin is extremely unlikely to be overturned at this point. It is in fact consolidating into a scientific consensus as key publications and new information or analyses are slowly diffusing through, mulled over, and unsuccessfully challenged by the wider scientific community. That includes Dr. Chan, who has long given up trying to bring her arguments to scientific journals and rather brings them to willing newspapers and politicians instead.

So instead of explaining (like many others did at length and in detail) why Dr. Chan’s various assertions in her previous New York Times op-ed are not getting any fresher despite a new NYT paint job, I wanted to add something new to the stale and repetitive lab leak discourse in 2026.

I will share results from my newest investigation into two different, and maybe more relevant origin questions that I have been chasing recently.

1. Can a scientifically coherent lab leak hypothesis still be formulated in 2026?

2. Has the lab leak theory been censored or was zoonosis the preferred narrative in the media?

Answering these questions will not only contextualize what to think about Dr. Chan’s recent New York Times opinion piece, but how we, as democratic citizens, have to prepare ourselves against being manipulated by those in power.

Sounds good? Let’s head in.

1) Can a scientifically coherent lab leak hypothesis still be formulated?

Let us get one thing out first: Whether SARS-CoV-2 originated in nature or in a lab is a scientifically testable question.

It is not a question of individual belief, human testimony, mystical explanations, secret intelligence, moral judgement or “other ways of knowing”. It is a question addressable within the domain of the physical world, where scientific evidence and inquiry are the only reliable method we have to test our hypotheses and not fool ourselves about objective reality.

That does not mean that we necessarily have sufficient evidence to prove or disprove a particular hypothesis. In science, proving the negative, that something did not happen, is usually tricky. On top of that, in outbreak investigations, one often operates with imperfect knowledge and data.

What makes a hypothesis scientific?

Scientists commonly think about it this way:

- How well does a proposed theory explain evidence and observations we already have?

- How well does a proposed theory predict the outcome of experiments and new observations?

A scientific theory also needs to observe parsimony. Which basically means that you reject unevidenced and unnecessary assumptions to make your "theory" or data fit.

The “lab leak theory” is commonly discussed as though it were a single hypothesis.

In reality, it is an amalgamation of origin claims that differ in mechanism, emergence route, in what would count as evidence for or against them, and in prior plausibility: deliberate construction of a bioweapon; de-novo genome assembly, gain-of-function research with genetic engineering or serial-passage adaptation of a bat virus; in vitro cultivation of a bat virus that escaped containment; or infection of a fieldworker or laboratory worker who then carried the virus to the city.

Some lab leak scenarios even propose researchers would have brought the virus from the lab specifically to the Huanan market, either directly or via discarded lab animals, making SARS-CoV-2 emergence from the market scientifically almost indistinguishable from zoonotic spillover from wildlife at the market brought in from vendors sourcing from the wildlife trading industry. Some propose wildly off timelines of emergence to, from military games in October or viral database takedowns in September to apparently many months earlier, as lab leak debater Peter Miller points out:

Peter Miller won an 18 hours, $100,000 lab leak debate against proponents by systematically working through many of their proposed scenarios and showing how they were incompatible with known facts and evidence.

Given the diversity of lab leak “hypotheses", many proponents of the lab leak theory either hold mutually contradictory view, decline to specify what they believe, or simply hope their claims are not taken through to their necessary implications where they would meet the body of known scientific evidence. Let’s illustrate with an example.

Did Dr. Fauci cause the pandemic?

This failure of taking a lab leak allegation to their necessary implications has been especially blatant recently, as US Senator Rand Paul (and as we will see, Dr. Alina Chan as well) are once again weaponizing lab leak allegations against retired public health servant Dr. Anthony Fauci; whom he accuses of having been involved in the creation of SARS-CoV-2 via gain-of-function research.

For Fauci to have any culpability in the origins of the pandemic, Senator Rand Pauls specific lab leak hypothesis necessitates two auxiliary assumptions:

• First, that SARS-CoV-2 is not a natural virus, but was somehow intentionally manipulated and came about through a gain-of-function research program.

• Second, that Dr. Fauci is somehow causally implicated in said research program, either through funding, knowledge of or outsourcing it to Wuhan.

Despite years of digging through official and private records, there is no evidence of any misconduct or culpability of Dr. Fauci. And yet, the problem Rand Paul’s allegations face do not just lie with the latter.

The real problem for Rand Paul allegations is that he is advocating for a particularly weak lab leak hypothesis that is contradicted by multiple lines of independently verified and undisputed scientific evidence.

SARS-CoV-2 has nothing to do with gain-of-function research or genetic engineering. We know this because any lab origin hypothesis that stipulates a non-natural or modified virus has very hard scientific evidence against it.

All engineering scenarios fail already at scientific analysis of the uncontested SARS-CoV-2 genome sequence itself that we have sequenced all around the world from millions of patients.

• SARS-CoV-2’s genome is a recombinant mosaic that could not be assembled in a lab.

• It houses a novel RBD motif that was found in bats only after the pandemic.

• It includes an FCS that is an out-of-frame insertion of a non-canonical motif, and with a superfluous, non-motif proline addition

• Mechanistic studies performed in 2022 discovered a new cleavage synergy with a proximal glycosylation and sequence motif, a hallmark of evolution, not design.

The furin cleavage site (FCS) in SARS-CoV-2 bears many hallmarks of evolution and not intelligent design. Furthermore, no designer could have created a hitherto unknown FCS-QTQTN-Glycan mechanistic synergy that was only discovered in 2022.

Neither gain-of-function research, nor genetic engineering, can credibly produce a virus with these features in any laboratory setup. The “gain-of-function research introducing an FCS via genetic engineering” hypothesis is among the weakest lab leak scenarios, albeit the most politically salient.

The goal of Rand Paul’s lab leak hypothesis is not parsimony or scientific plausibility, but getting Dr. Fauci.

Now one problem I have with most journalists covering this origin topic is that they are always eager to cover the high-profile dispute, parse whether they can find any substantiation of Rand Paul’s allegations in selectively disclosed and quote-mined documents; but that they never bother to even differentiate what lab leak scenario these proponents actually endorse.

This is journalistic malpractice.

There are other lab leak hypotheses that can account for some of the genetic and mechanistic features that any engineering hypothesis can not. For example, a natural bat virus could have been collected and cultured that already had the right genetic setup to infect humans. Or a serial passage experiment in susceptible animals was performed on such a natural virus (which does not make any sense to perform on a novel, unknown virus; but just for arguments sake).

Now of course, other lines of evidence exist that have been collected by independent researchers, methodologies and studies that make these lab leak hypotheses also hard, if not impossible, to square with the body of verified scientific evidence; such as case and spatial epidemiology, phylogenetic analyses and lineage tracing, and metagenomics of environmental samples taken from the Huanan market. Below I created a constraint sieve chart visualizing how accumulated evidence over the years makes it actually quite impossible to formulate a lab leak hypothesis that can explain the totality of independent evidence we possess today, from genomic features and mechanistic studies bearing hallmarks of evolution, to phylogenetic history showing two introductions, to case and spatial epidemiology centered around the Huanan market, to metagenomics and transcriptomic samples placing SARS-CoV-2 susceptible and sick animals inside the market.

Constraint sieve chart of commonly proposed lab leak hypotheses and how well they can account for various published lines of evidence.

“We dont know what we dont know” or “Chinese obfuscation” only work as arguments to keep the lab leak idea in play if the available evidence is poor.

But the available evidence to assess the origin question is not poor anymore, and has not been for quite some years now.

The power of the zoonotic origin theory does not rely on a particular piece of evidence or study, nor does a single line of evidence or inquiry have the power to disprove any and all lab leak hypothesis.

Real evidentiary power in science arises through triangulation; where multiple independent lines of evidence, collected by independent researchers, with independent methodologies, are all readily explained by one parsimonious scientific theory.

The reality is that given the available evidence today, one is strained to even formulate any new lab leak hypothesis that is scientific in the operative sense that it can explain the available data and maintain parsimony.

We outlined that a scientific theory not only needs to be able to accommodate existing evidence parsimoniously, but ideally have predictive value that can be confirmed via new discovery or experimentation.

But we are not just starting the discussion now, the lab leak theory in its many iterations has been around ever since the virus took Wuhan, and then the rest of the world, by storm.

What was the predictive power of the competing origin hypotheses since 2020?

Both "zoonotic spillover at market" & "lab leak from WIV" theories were offered immediately as alternative origin explanation in early 2020. Mired in initial uncertainty, a scientifically reasonable version of the lab leak hypothesis could be postulated without running into evidentiary contradictions.

Given the constant goal-post shifting and diversification of origin speculations to account for ever increasing contradictory evidence, I think it can be a useful exercise to retrospectively check of how accurately both theories predicted subsequent research findings that took years to accumulate, or failed to stack up for that matter.

Evidence-prediction matrix for whether each proposed hypothesis predicts an subsequent discovery or observation, merely accommodates it, is neutral toward it, or stands in tension with it. “Consistent*” denotes consistency achieved only through an added, unevidenced assumption.

While the above matrix is neither comprehensive not free from some personal assessment, I believe it is fair to state that the zoonotic spillover at market theory has predicted subsequent findings extraordinarily well and remains virtually unchanged since its first formulation in 2020.

In stark contrast, the lab leak theory has constantly shifted speculations, trying to justify its existence against contradictory evidence. It predicted nothing about what subsequent studies would find or discover since 2020.

So when a theory has shown no predictive power, does not observe parsimony, and can not even be formulated in a way that is consistent with available evidence, why are we still pretending this is a scientific controversy around competing hypotheses?

There is no scientific lab leak theory or hypothesis any more. That is why I have for some time now argued that we should stop calling the lab leak a theory, or even a "conspiracy theory".

It is something different; an emotive, engaging narrative that mutates over time and according to the preferences of who is in power. The Rand Paul version currently leading the news cycle is just its latest, most inflammatory, iteration.

What explains the endurance of this lab leak narrative against a scientifically unchallenged zoonotic origin explanation?

This brings us to our next part:

2) Has the lab leak theory been censored / was zoonosis the preferred narrative in the media?

For many years, but latest since 2020, US Republicans and associated media operators and influencers have been building up the myth of media bias and censorship against American conservatives. (subsequent studies find the opposite)

Triggered by the big lie of a supposedly stolen 2020 presidential election, social media platform moderation teams, mainstream media fact checkers, and disinformation researchers like the brilliant Renee DiResta have become a thorn in the eyes of the burgeoning fascists MAGA movement in power today.

Between the supposed censorship of the Hunter Biden Laptop story of 2020 and the manufacture of the Twitter-files “censorship industrial complex” narrative in 2022, there was one particular topic where the rightwing propaganda machine had set out to finally break the social contract of being held accountable, or at least fact checked and counter-spoken to, when spewing lies and conspiracy theories.

You guessed right. The supposed suppression of the lab leak theory.

As often, propaganda is most effective when there is a kernel of truth, a key fact or anecdote that can be retreated to when challenged. In this case, Facebook has indeed moderated assertions of a man-made virus on its platform between February-May of 2021.

To give some context:

And in the aftermath January 6th 2021, where disinformation has brought a violent mob to storm congress in order to overturn a democratic election, social media platforms scrambled to implement stricter content moderation and enforcement of their own stated community guidelines. Especially Facebook and Twitter found themselves in hot water for enabling these insurrectionist actors and act; bound to be facing potential liability or at least increased scrutiny by lawmakers that almost have been lynched. So they responded quickly by tightening moderation enforcement and deplatforming the biggest offenders. Most visibly, then president Trump, who had incited the storm on the capitol, was suspended from Twitter. A catastrophic event for those of the rightwing propaganda machine whose power and fortune relied on continuously lying to their audiences.

Stolen election and other conspiracy theories associated with the MAGA movement quickly came under scrutiny too; and this is where I assume Facebook made that call on “explicitly prohibited and removed content stating that "COVID-19 is man-made or manufactured" under its COVID-19 misinformation policy.

Worth knowing that in 2020, the most dominant lab leak theory was its bioweapon version conjured up by Steve Bannon and Miles Guo, Trump operatives strongly tied to the MAGA ecosystem. Additionally, Miles Guo’s scam operation led to defrauding believers in a man-made virus with financial scams amounting to a billion dollars.

One can reasonably agree or disagree with Facebook’s temporary moderation decision, but not with the fact that Facebook is a private company who made this decision unilaterally, not the government or the media at large. Twitter for example never censored lab leak content. Neither did Reddit, or YouTube, or the mainstream media, including what was once called legacy press or quality newspapers.

Nevertheless, once Facebook reversed the policy a few months later, the rightwing propaganda machine seized upon the moment to claim that the “conspiracy theorists” were right after all. Subsequently, the lab leak theory became totemic as a symbol for censorship overreach and the suppression of conservative views and voices in media at large.

Most in mainstream media and elite circles either fell for the propaganda or decided it was necessary to signal their own performative neutrality on the lab leak question, and soon, willing op-eds started appearing across the quality and mainstream press, with articles such as:

The last-week-of-May 2021 wave of press self-criticism over the supposed dismissal of the Covid-19 lab-leak theory. Opinion columns and newsletter essays published 24–31 May 2021 arguing that mainstream news outlets had wrongly dismissed the hypothesis that SARS-CoV-2 escaped from the Wuhan Institute of Virology, or had suppressed legitimate inquiry into it.

It became common vibe-based knowledge that the lab leak theory was suppressed by fact checkers and scientists, and had been dismissed by the media. Some have argued these events also changed the lab leak narrative from a low-status belief into a high-status one.

In turn, the success of the supposed lab leak theory redemption arc just boosted the hunger of the rightwing propaganda machine to get rid of any and all content moderation, and as we learned later, any checks on their power to lie without consequence and accountability.

But what if the vibe was wrong?

I have spent years trying to reconstruct the major events that shaped public perception of the origin controversy, and I came to the conclusion that while legitimate uncertainty existed, especially in early 2020, everything that unfolded after can not be explained by scientific ambiguity or Chinese obfuscation alone.

The gap between scientific evidence and what scientists believe, versus what the public, including elites in media, politics and business believe, is among the biggest for any scientific topic in existence, including climate change.

Where does the striking evidence-belief gap come from? What prompts the New York Times to again and again platform lab leak activists like Dr. Alina Chan, or resident lab leak columnist Zeynep Tufekci for that matter?

Is it for performative neutrality? To balance reporting by giving baseless insinuations and unsubstantiated allegations equal weight to a body of scientific evidence? To fend off charges of bias or suppression? To remedy what is assumed true, that the lab leak theory did not get a thorough hearing by the media and in public discourse? I can not answer these questions for the NYT.

What I can do, however, test whether the lab leak theory has indeed been unfairly treated in the media.

For this novel analysis, I will use two data sets that can be obtained freely without special API access to platforms. The first dataset is from MediaCloud, an open-source platform for media analysis. The second is YouTube Data tools, a front end to the YouTube Data API search.

Both allowed me to search for and assemble a comprehensive database of online-news articles and Youtube videos talking about the origin controversy published between Jan. 1 2020 and July 31st 2026. (please see method section at the end of the article for further details).

Using the context-aware natural language processing of advanced large language models (Claude Opus 4.8, Claude Opus 5.0), we assessed 11,086 COVID-19-origin online-news headlines and 1,546 YouTube videos on a mirrored seven-point ordinal scale from −3 (natural/zoonotic origin asserted as fact) through 0 (neutral, uncertain, both-sides or procedural) to +3 (laboratory origin asserted as fact), with graded intermediate points.

On the news articles dataset, I find that the lab leak theory has never not been the dominant theory in online news media, with the exception of the first quarter of 2020, where the emerging pandemic fueled a lot of reporting on wildlife trade and explainers of viral emergence through zoonosis.

After that, starting latest with the April-June of 2020 quarter, the Trump administration and MAGA movement has fully embraced the “China virus” narrative, columnist Josh Rogin sold a BS story of secret diplomatic cables to the Washington Post about a supposed biosafety warning from a diplomatic WIV visit, while conspiracy theorists like Yuri Deigin kicked-off gain-of-function fearmongering, DRASTIC began selling the dead Mojiang miner story and Dr. Alina Chan did her first front-page media tour based on a flawed “pre-adaptation” preprint that still has never managed to pass peer review.

From around April 2020 onward, the lab leak theory never left the public discourse, specifically on the right and partisan news media which also dominate in overall article volume at least 3:1 over centrist and left media.

Per-quarter distribution across the seven-point stance scale, full dataset (left) versus US Top 50 Newspapers (right), each bar normalised to 100%. Low-volume quarters produce spiky distributions and should not be over-read.

That lab leak dominance pattern holds largely true even when we remove the rightwing partisan news media and just focus on “Quality press” or legacy news media, which gave the lab leak theory plenty of endorsing coverage. The biggest difference between rightwing partisan and legacy media is that legacy media also reported on new scientific discoveries when they happened (e.g market origin paper from Worobey et al. and Pekar et al. in Q2 of 2022, and raccoon dog-metagenomics paper from Crits-Christoph et al. in Q3 of 2024).

Yet additional evidence for zoonosis never seemed to last in whether to assess the origin question one way or the other, and soon even legacy media would be back with favorable lab leak coverage.

One illustrative, albeit not perfectly representative, example of lab leak coverage reflects the trajectory of Fox News and NYT over the years. While Fox news always followed a hard lab-leak endorsing stance, the NYTs had infighting between their science and reporting desks versus their op-ed and political desks. With the slow pace of scientific findings, but accelerated GOP witchhunts of recent years, the NYT got pulled stronger and stronger to endorse the lab leak theory, a stance it clearly continues in 2026.

Origin coverage lean of NYT versus Fox News between 2020-2025

We can also have a look at the aggregate of lab leak partisan coverage from outlets that are either right-leaning, center or left-leaning.

Here are the top-10 origin coverage providers in US media for each bucket, with the number of origin-related headlines, lab: zoo coverage ratio, average endorsement score and confidence intervals.

Origin coverage volume and endorsement stance per outlet political lean.

What is striking here is the dramatic difference in volume of rightwing media, and how one sided they present the lab leak theory to their viewers. On average, for every 14 articles endorsing a lab leak theory, only 1 would speak supportively of the zoonotic theory. (and honestly, some of that might be LLM attribution error for ironic articles). So this is being generous. For right-leaning media, taking no position at all or both-siding the alternative theories is also not very common.

This is in stark contrast to how the NYT and other centrist media clearly operate, where every story where there are two sides deserve equal opportunity and coverage, leading to a lot of 50:50 coverage or tit-for-tat (e.g news desk reported a scientific discovery for zoonosis; political/op-ed desk runs an op-ed on how a lab leak can not be discarded given the new discovery).

Another interesting result is that left-leaning media, despite often being presented or decried by centrists as equally bad than rightwing media, are barely edging out a slightly more reality-aligned lab leak coverage than their centrist peers. So if I had to summarize:

• Rightwing and partisan news media dominate online news coverage on the lab leak topic and it has been entirely one-sided with a strong slant towards endorsement.

• Centrist news media can somewhat be trusted in their scientific reporting, but not in their editorial stance, selection of what to cover, and endorsement.

• Left-leaning news media just plays a strongly diminished role in public discourse of the origin question, despite often more accurate coverage than centrist media.

Now at the moment, we looked at news article volumes and endorsement, but just pushing out more and more digital content like rightwing propaganda outlets tend to do does not necessarily give a good indication of the consumption or popularity of that content.

So I wanted to have at least one dataset that has real engagement data, and luckily for YouTube videos, this was easily obtainable.

So next, I used the same LLM-based methodological approach and criteria to rate the lean of YouTube videos. For anybody who has held out hope that algorithmically driven entertainment platforms would somehow be more accurate or at least diverse in their lab leak coverage, I am sorry to disappoint.

View-weighted stance of 1,546 on-topic YouTube videos on SARS-CoV-2 origins (only videos with more than 1,000 views) from 2020–2026. Each dot is a video and its area is proportional to view count; the most-viewed videos are labelled. Positive values lean pro-lab-leak. The unweighted mean is +0.62; weighting by views raises it to +1.08.

Scientifically accurate or aligned pro-zoonotic YouTube content dramatically underperforms not only on creator volume, but also on engagement itself.

While only about 14% of all YouTube videos on the topic strongly endorse the lab leak theory, they win 36% of all eyeballs for the topic.

43% of all YouTube origin videos show a clear lean towards the lab leak theory, and gather the largest portion of attention with 41% of overall views.

In contrast, of the 7% of all YouTube videos that either lean or strongly endorse the zoonotic origin theory (including multiple virologist interviews of ours) only manage to gain 2% of all attention for the topic.

The problem with engagement-algorithm driven systems is that the winner usually takes all; so the distribution of attention is very asymmetrical towards a few highly influential videos pro-lab leak videos among a cottage industry of conspiratorial video creators. You can explore the YouTube data interactively here to form your own impression too. Let me know what you find.

I think the classic example for the lab leak theory is when late-night comedian Jon Stewart endorsed the lab leak theory on the late night show with Stephen Colbert, seen by 8.6 million people as of today.

Most popular origin-themed YouTube videos receiving over 1 million views.

Now one caveat of the YouTube dataset is that the videos need still to be live in order to be recovered, so some influential videos such as when Tucker Carlson platformed Steve Bannon’s bioweapon whistleblower Dr. Li Meng Yan in 2020 that received 8.8 million views have since been removed.

So the endorsement and engagement slant of the YouTube lab leak dataset is possibly an underestimate, specifically around conspiratorial content that might have been removed since 2020.

So let us sum up the analysis:

We observe that origin coverage in US digital news media, including US top 50 newspapers, as well as on the video platform YouTube significantly leans toward endorsement of the lab leak hypothesis across the investigate period from 2020 until 2026, and the observed endorsement strengthens sharply with audience reach.

Partisan and mainstream outlets that endorse the lab leak theory dominate news volume, and the majority of the most popular “Covid origin” YouTube videos exceeding one million views endorses the lab leak theory, while none endorses a zoonotic explanation.

Except for Q1 of 2020, where viral explainers and neutral pandemic coverage touched on viral emergence, there has been no subsequent period in time where origin media coverage even approximately reflected the state of scientific evidence and the corresponding likelihood of zoonotic versus a laboratory origin explanation.

The vibe was wrong. The lab leak theory was the preferred narrative all along and remains so today.

The lab leak theory was not suppressed nor discarded. It has been the dominant theory promoted by politicians in the highest offices, partisan and mainstream news outlets, legacy and alternative news media; and some of the most popular media personalities on the left and right.

The common knowledge that one could not even talk about the lab leak theory without serious repercussions was mostly a supercharged fabrication by a partisan media machinery based on a true, but not representative, anecdote of a Facebook moderation decision between Feb-May 2021. Mainstream media and elite belief accepted it as truism for various reasons, turned endorsement of lab leak theory into a high-status belief, while the public has been sold the myth of the lab leak theory suppression ever since. This is astonishing, given that in 2026, there is no scientifically legible lab leak hypothesis left to defend, and there is also no professional upside to pointing that out. The lab leak emperor is naked, and everybody making editorial decisions, especially the Grey Lady, just pretends to see the most compelling ball gown.

Ironically, it appears to me that a type of media group think did happen for a considerable fraction of legacy media. When a belief becomes costly to contradict in public, journalists and editors stop contradicting it in public, and their silence is then read as agreement by colleagues, which raises the cost further. Editorially discarding the lab leak theory as unevidenced, or even a baseless conspiracy theory, has become socially and professionally costly, for individual journalists, commentators and outlets alike, reinforcing the continued astonishing pattern of media coverage bias.

When was the last time you read in the New York Times that the lab leak theory is a baseless conspiracy myth; which is much closer to objective reality than whatever Dr. Chan cooked up in her recent op-ed?

So what explains the evidence-belief gap between what scientists know and the public believe on the origins of SARS-CoV-2?

It is certainly not scientific ambiguity.

If I had to propose a hypothesis, the astonishing and ongoing media coverage bias across news, digital and social media would certainly be a prime candidate for further investigation.

What now?

One argument I make in my book is that in the information age, where technological disruption has upended the news media business and fundamentally reshaped incentives, coverage tends to follows elite belief, not factual reality. From oligarchic consolidation of media ownership, partisan operatives, a business- and politician-friendly press corps to client propagandism for billionaire interests, extreme wealth found in the precarious economic situation of technologically disrupted media businesses an opportune moment to acquire and enshrine their power over society.

Science, like journalism, performs an important function for democratic society. It challenges narratives of power with fact. Facts establish a relationship to objective reality, to truth. Only where truth exists can power be held accountable.

Yet in this moment in the US and other places, the scientific community finds itself increasingly at odds with the beliefs of the powerful, and that has many unwelcome consequences.

Scientists in the US are attacked. Losing their funding, jobs, career prospects or livelihoods. If they have any public profile, they might get targeted by hate mobs, domestic terrorists, and increasingly a weaponized US government; endangering their physical safety and freedom. If you have not heard much about this, you should consider reading this:

Universities and federal scientific institutions are similarly under siege, threatened by lawfare, starved of resources and funding, subjected to new, often unlawful or unconstitutional rules, arbitrary enforcement and political interference.

One consequence of being at odds with power that is not often talked about is the permanent drumbeat of partisan and mainstream media, alternative and popular content creators, podcasters and client propagandists who are financially incentivized and politically rewarded to justify and foment such attacks on scientists and universities.

The fascist media machinery is of course the most visible driver of false narratives like the lab leak theory and needs to be destroyed for democracy to thrive. There is no way around that.

But we also need deeper media reform beyond that. It has been particularly disappointing when even influential centrist media, famous for observing performative neutrality, breaks their own neutrality rules with lab leak-favorable coverage dominating almost 2:1, with a mean slightly above the “both theories are equally likely” and a tilt towards “but if I had to guess the lab looks sus, doesn’t it?”

Particular scorn goes towards the economically thriving Grey Lady that basically enjoys a monopoly in journalistic influence after the disruption of the market by technology. They are the ones that have the resources to not cater to power, yet choose to do so. Even outside their usual performative neutrality stance on topics like “whether actual fascists or people calling fascists them deserve scolding?”, the The New York Times is very clearly and consistently taking an editorial position towards endorsing and promoting the lab leak theory that fueled and still fuels a lot manipulative attacks on science. There is no way to sugarcoat this, and if you do not want to hear it from me, see what Michael Mann and Peter Hotez have to say about the Grey Lady in their book on anti-science.

Platforming the known lab leak activist Dr. Alina Chan, again, to fearmonger about gain-of-function research and allocate blame onto innocent scientists, while the news cycle predictably jumped on the Senate hearing of Dr. Fauci, is not “without fear or favor” coverage.

It is a dereliction of duty to objective reality, and as partisan and obedient to power as one can be.

In a world where power can turn lies into reality, we all need a functioning fourth estate more than ever. There are many excellent journalists working at the New York Times, it is a large organization. But if the Times does not have a reckoning on its editorial and leadership’s cozy relationship with narratives of power, something new and journalistically independent will have to take its dominant place.

In the meantime, readers are advised to exercise scepticism and withhold trust. Especially when NYT spends their vast resources on producing a glitzy op-ed instead of an investigative story, we should understand what faction of the times has won the editorial greenlight, and why.

Methods section

News-headline corpus construction

Headlines were drawn from the Media Cloud online-news archive across four high-reach United States collections (US Most Visited News Online [Mar 2025], U.S. Top Sources 2018, U.S. Top Newspapers 2018, and United States - National) for 1 January 2020 to 31 July 2026. To capture broad coverage without directional bias, articles were retrieved using a disjunctive Boolean query combining neutral baseline terms (e.g., “covid origin”, “viral origin”) with frame-specific keywords balanced across competing hypotheses: natural spillover (e.g., “wet market”, “zoonotic”, “pangolin”, “raccoon dog”, “Huanan”, “spillover”), laboratory accident and research funding (e.g., “lab leak”, “Wuhan Institute of Virology”, “gain of function”, “EcoHealth”), and engineered origins (e.g., “bioweapon”, “man-made”).

This search returned 92,255 English-language records across 247 outlets after de-duplication on record ID. The unit of analysis is the headline. Because the collections are reach-weighted to reflect broader news consumption reality of US digital news media consumers, the online news corpus over-represents high-traffic partisan and tabloid outlets relative to legacy newsrooms. We analyze full news media coverage as well as the subset of top 50 US newspapers coverage to track similarities and difference of origin coverage between these news formats.

To establish on-topic status of news headlines — whether an article concerns the origin of SARS-CoV-2 rather than COVID case counts, policy or unrelated health news — was set by a relevance classifier: a TF-IDF plus logistic-regression model trained on 10,583 positive-labelled IDs, thresholded at 0.74. Against a 140-item hand-labelled gold test held out from training, the classifier reached precision 0.985, recall 0.944 (F1 0.964), outperforming a rule-based keyword baseline (precision 0.958, recall 0.972) and a simpler keyword baseline (precision 0.979, recall 0.648) evaluated on the same set. Applied to the full corpus, the classifier’s on-topic set (10,577 headlines) was combined by union with an independently assembled reference set (10,583 headlines), giving a final on-topic corpus of 11,082 headlines (overlap 10,078; Jaccard 0.909 between the two sets). This union was validated against independent human coding on a stratified 220-item sample (seven strata by predicted probability and source): union precision was 0.933 (95% CI 0.868–0.999) and the off-topic false-negative rate was 3.49% (95% CI 1.73–5.24%), implying a corpus-wide true on-topic count of approximately 13,177 headlines (95% CI 11,576–14,779) — the reported on-topic corpus is a conservative, high-precision subset rather than an exhaustive one. Agreement between the two independent coders on this sample was Cohen’s κ = 0.890 (9 of 220 disagreements), with essentially all disagreements occurring on headlines that require real-world context not recoverable from the headline text alone.

3.2 Stance instrument

Each on-topic item was scored on a mirrored seven-point ordinal scale from −3 (natural/zoonotic origin asserted as fact) through 0 (neutral, uncertain, both-sides or procedural) to +3 (laboratory origin asserted as fact), with graded intermediate points. The “lab-lean” reported throughout is the mean on this scale, positive values denoting a lab-ward tilt. All 11,082 on-topic headlines were scored by a single large-language-model rater (Claude Opus 4.8, Anthropic), processed sequentially in batches with an explicit, cumulative set of calibration anchors (e.g., named-official assertive lab leak claims scored ±2, procedural WHO-investigation headlines scored 0; a specific worked example — “WHO panel says lab leak theory extremely unlikely, not worth exploring further” — anchored a canonical +2 zoonosis rating) to hold scoring consistent across the full run. Before analysis, 971 exact title-and-outlet duplicates (predominantly wire-service reposts of a single story) were removed, and a further 632 headlines were flagged as not actually bearing on the origin question on inspection.

YouTube corpus construction and coding

Videos were retrieved with YouTube Data Tools, a front end to the YouTube Data API search endpoint; because that endpoint caps a single query at roughly 500 results, the corpus was assembled as a succession of day-by-day searches to approximate a census, yielding 15,611 videos published January 2020–July 2026. Restricting to videos with more than 1,000 views (9,296 after de-duplication) and then to origin-relevant content left 1,546 on-topic videos, every one rated by Claude Opus 4.8 on the same seven-point scale. Because each record carries a cumulative view count, both an unweighted and a view-weighted mean are reported, the latter approximating the stance of the average viewing rather than the average upload. Reliability was assessed against an independent rater on a blind 200-video sample (100 pipeline-on-topic, 100 discarded): on/off-topic precision 0.98 and recall 0.87 (weighted to the true prevalence); on the shared scale, agreement within one point for 94% of videos, quadratic-weighted κ = 0.73, and a mean human-minus-study difference of +0.20 — the rater leaned marginally more lab-ward, so the reported tilt is conservative.