A comprehensive fact-check and literature review exposes not just factual inaccuracies but fundamental logical paradoxes that undermine the article’s core message about “black swan events” and vaccine-variant mismatches.
INGA314.ai analysis
https://jamanetwork.com/journals/jama/fullarticle/2837354
The August 2025 JAMA article “SARS-CoV-2 Variants Challenge Surveillance Efforts and COVID-19 Vaccines” presents a narrative of viral unpredictability punctuated by confident predictions about future threats. This apparent contradiction—claiming variants are unpredictable while making specific forecasts—reveals deeper tensions in how we communicate scientific uncertainty during public health crises. A comprehensive fact-check and literature review exposes not just factual inaccuracies but fundamental logical paradoxes that undermine the article’s core message about “black swan events” and vaccine-variant mismatches.
The article’s historical timeline claims prove largely accurate: Omicron did emerge in November 2021, JN.1 appeared in late 2023, and vaccine formulations evolved from 2022’s bivalent vaccines through 2023’s XBB.1.5 targeting to 2024-2025’s KP.2 focus. However, the interpretive framework surrounding these facts—particularly the characterization of variants as “black swan events” and the framing of vaccine-variant relationships as problematic “mismatches”—contradicts established scientific understanding. The COVID-19 pandemic itself fails to meet Nassim Taleb’s criteria for black swan events, as numerous experts had warned about pandemic risks for decades. Even Taleb himself called COVID-19 a “white swan”—entirely predictable given known risk factors.
The scientific literature reveals that SARS-CoV-2 evolution contains both predictable elements (mutation accumulation rates, selection pressures, convergent evolution) and genuinely surprising aspects (timing of emergence, specific mutation constellations, magnitude of evolutionary jumps like Omicron’s 30+ spike mutations). This nuanced reality gets lost when authors claim complete unpredictability while simultaneously making confident predictions about future variant behavior—a logical inconsistency that pervades not just this article but much pandemic discourse.
The paradox of predicting the unpredictable
The most striking logical flaw emerges from what researchers term “the paradox of scientific advice”—the impossible demand that scientists remain neutral while providing useful guidance. Studies demonstrate that experts routinely claim SARS-CoV-2 evolution is fundamentally unpredictable while making specific predictions about variant behavior. Massachusetts General Hospital’s model predicted “over 23,000 deaths within a month of Georgia reopening” when actual deaths numbered 896, exemplifying how confident predictions coexist with claims of uncertainty. This selective certainty—expressing high confidence about potential dangers while claiming uncertainty about positive outcomes—suggests strategic rather than epistemological positioning.
Recent research published in Nature demonstrates prediction models achieving “high accuracy (area under the receiver operator characteristic curve = 0.92 to 0.97)” for variant mutations up to four months in advance. These successes directly contradict blanket unpredictability claims. The disconnect between laboratory confidence and real-world complexity creates what philosophers call the “Large World vs. Small World confusion”—treating complex evolutionary dynamics with fundamental uncertainty as if they were statistical problems with known probabilities. In vitro studies use definitive language (“will evade,” “significantly reduces”) based on limited experimental conditions, yet field studies consistently show more nuanced outcomes than laboratory predictions suggest.
The temporal dimension adds another layer of contradiction. Prediction accuracy degrades dramatically beyond 2-4 weeks, yet policy recommendations routinely rely on projections extending months into the future. Models claiming 10-week accuracy report mean errors of 9%, masking enormous variability and systematic biases. Expert consensus identifies meaningful prediction windows measured in weeks, not months or years, making claims about patterns over “5-6 years” scientifically dubious.
Fact-checking the “black swan” narrative
The article’s claim of “three black swan events in 5-6 years” warrants careful scrutiny against scientific definitions. True black swan events must be outliers beyond regular expectations, carry extreme impact, and appear predictable only in hindsight. The COVID-19 pandemic itself fails all three criteria: multiple warnings existed, precedents included SARS and MERS, and many experts considered a respiratory pandemic inevitable rather than surprising. While Omicron’s emergence with over 30 spike mutations genuinely surprised scientists, characterizing expected challenges as “black swans” misrepresents normal pandemic dynamics.
The XBB variant’s designation as a “black swan event” proves particularly problematic. Scientific literature shows mixed views on this classification. XBB emerged through recombination of two BA.2 lineages during summer 2022—unusual for SARS-CoV-2 but not unprecedented in coronavirus evolution. WHO’s Technical Advisory Group initially declined to designate XBB variants as separate variants of concern, suggesting they fell within expected evolutionary patterns. Some researchers argued the recombination was “not entirely unpredictable” based on known viral evolution mechanisms.
The vaccine formulation timeline the article presents is accurate: FDA authorized bivalent vaccines targeting original strain plus BA.4/BA.5 on August 31, 2022; approved XBB.1.5-targeting vaccines on September 11, 2023; and approved KP.2-targeting vaccines in August 2024. However, framing this as a problematic “mismatch pattern” misrepresents the normal, expected challenges of vaccine development against rapidly evolving pathogens. Multiple sources confirm vaccines maintained effectiveness against severe disease despite viral evolution—the system worked as designed, not as a failure.
Survivorship bias distorts variant perception
A critical flaw in variant analysis involves what statisticians call survivorship bias—focusing exclusively on successful variants while ignoring failures. Scientific literature disproportionately examines variants that achieved dominance (Alpha, Delta, Omicron) while the “graveyard” of variants that emerged but didn’t spread remains largely unstudied. This creates false impressions of predictable success patterns and inflates apparent prediction accuracy. For every Omicron that surprises with rapid global spread, dozens of variants with concerning mutations fail to establish themselves—but these failures receive minimal analysis.
Wastewater surveillance reveals “cryptic SARS-CoV-2 lineages” circulating undetected by clinical sequencing, suggesting our picture of viral evolution remains incomplete. Only 78% of high-income countries sequence more than 0.5% of cases, while just 42% of low- and middle-income countries reach this threshold. Geographic surveillance gaps mean variants could circulate at up to 21.7% prevalence before detection in low-capacity systems. These blind spots make claims about comprehensive understanding of viral evolution patterns premature.
The focus on dramatic outcomes creates what researchers term “publication and reporting bias.” Successful predictions receive disproportionate attention in scientific literature and media coverage, while failed predictions get buried in technical appendices or remain unpublished entirely. This systematic bias inflates apparent prediction accuracy and obscures true uncertainty levels, creating overconfidence in our ability to anticipate viral evolution.
Cross-protection complexities undermine simple narratives
The article’s framing of vaccine-variant relationships as problematic “mismatches” oversimplifies complex immunological realities. Research demonstrates that immunity involves multiple mechanisms beyond neutralizing antibodies that the article implicitly prioritizes. CD8+ T cells, essential for controlling severe infections, show greater cross-reactivity between variants than antibody responses. Most T-cell epitopes remain conserved across variants, providing durable protection even when antibody neutralization wanes.
Real-world evidence contradicts simple mismatch narratives. Previous BA.1/BA.2 infection provides strong protection against BA.5, while “double-primed” individuals with both non-Omicron and Omicron exposures show 48% lower reinfection rates. The 2024-2025 vaccines demonstrate 33% effectiveness against emergency department visits and 45-46% against hospitalizations—significant protection despite imperfect variant matching. This represents expected vaccine performance against evolving pathogens, not failure.
The scientific consensus recognizes that respiratory tract immunity from previous infection or vaccination provides cross-protection against distinct variants “in the absence of variant spike-specific neutralizing antibodies.” Non-antibody immune mechanisms including T cells and innate immunity play crucial roles often overlooked in simplified discussions of variant escape. While limitations exist—particularly differences between systemic and mucosal immunity—the overall picture shows remarkably robust cross-protection mechanisms.
Why pandemic predictions fail: institutional and cognitive factors
The systematic production of logical inconsistencies stems from structural incentives within scientific and public health institutions. Publication bias rewards dramatic predictions and successful forecasts while burying failures. Political pressure demands definitive guidance that conflicts with genuine uncertainty. Professional reputation incentives push scientists toward appearing confident and authoritative even when evidence remains limited. These factors combine with cognitive biases—overconfidence in predictive accuracy, confirmation bias favoring supportive evidence, and hindsight bias making successful predictions appear obvious retrospectively.
Communication constraints amplify these problems. Complex uncertainty must be reduced to actionable guidance for public consumption. Scientists face the impossible task of maintaining authority while acknowledging limitations. Any admission of uncertainty risks political weaponization by critics. This creates what researchers call “the certainty trap”—scientists become more confident in public statements than their private assessments warrant, eliminating natural self-correction mechanisms that uncertainty acknowledgment provides.
The mask guidance reversal of early 2020 exemplifies these dynamics. Initial claims that masks were “absolutely” ineffective reflected supply concerns presented as scientific conclusions—a hidden value judgment masquerading as neutral fact. The failure to acknowledge uncertainty made subsequent changes appear arbitrary, undermining public trust. Similar patterns emerged in vaccine allocation debates, where equity arguments were presented as epidemiological conclusions, and reopening predictions that confidently forecast complex social-epidemiological interactions months in advance.
Toward epistemic humility in pandemic science
The analysis reveals that logical inconsistencies in variant unpredictability claims reflect deeper tensions between democratic needs for authoritative guidance and the intrinsic uncertainty of complex evolutionary systems. Rather than resolving these tensions through false confidence, the path forward requires what scholars term “epistemic humility”—recognizing that certain aspects of complex systems may remain fundamentally unpredictable.
Effective pandemic preparedness demands distinguishing between different types of uncertainty. Some unknowns represent temporary knowledge gaps that research can fill. Others reflect fundamental limits on predictability in chaotic systems. Conflating these categories leads to overconfident predictions that ultimately undermine scientific credibility. Building public understanding that scientific guidance appropriately evolves with evidence—rather than representing failure—becomes crucial for maintaining trust during future health crises.
The COVID-19 pandemic exposed the urgent need for what researchers call “a new epistemology of science for the public good”—one that navigates irreducible uncertainty while maintaining public trust and enabling effective decision-making. This means embracing process epistemology that recognizes viral evolution as a dynamic system resistant to static predictive models. It requires systematic study of failed variants and unsuccessful predictions to understand true base rates of predictive success. Most importantly, it demands transparent communication about when recommendations involve value judgments rather than pure scientific conclusions.
Conclusion: embracing uncertainty as strength, not weakness
The JAMA article’s core factual claims about variant emergence timelines and vaccine formulations prove largely accurate, but its interpretive framework—characterizing expected evolutionary challenges as unprecedented “black swan events” and normal vaccine adaptation as problematic “mismatches”—contradicts established scientific understanding. These mischaracterizations matter because they shape public perception and policy responses to ongoing pandemic challenges.
The deeper lesson involves recognizing that acknowledging uncertainty represents scientific strength, not weakness. Viral evolution contains both predictable and unpredictable elements, and pretending otherwise through false confidence ultimately undermines pandemic preparedness. The most honest scientific communication admits that while we can identify general patterns and pressures, specific evolutionary trajectories remain partially opaque to prediction. This reality makes robust surveillance systems, adaptive vaccine platforms, and maintenance of public health infrastructure essential—not because we can predict the next variant, but precisely because we cannot.
Moving forward requires abandoning the pretense that science can eliminate uncertainty about complex biological systems. Instead, we must build resilience through systems designed to function despite incomplete knowledge. The choice is not between false certainty and paralysis, but between honest acknowledgment of limitations and productive action within those constraints. In this light, the JAMA article’s logical contradictions serve as a cautionary tale about the dangers of conflating scientific authority with predictive omniscience—a confusion that ultimately serves neither science nor public health.
System D 