Robert F. Kennedy Jr.'s Autism Fictions

I just read Hannah Arendt's essays "Truth and Politics" and "Lying in Politics" for a Continental philosophy reading group I recently joined. In "Lying in Politics", what troubled Arendt was not simply that officials lie, but that they construct entire alternative realities, elaborate "images" that displace factual truth altogether. She distinguished between traditional lying, in which the liar knows the truth and the deception remains limited, and modern organized lying, in which collective manipulation becomes so pervasive that even the liars can no longer know what the truth is. The Pentagon Papers revealed something more unsettling than cynical deception: the architects of the Vietnam War, the celebrated "problem-solvers" and technocrats, had become trapped within their own fabrications. They may no longer have been able to distinguish between the world they had invented and the world as it was. They lied not for their country's survival, which was never truly at stake, but for its "image."

Her concept of "image-making" as different from traditional political deception might be helpful for understanding a certain contemporary phenomenon. Consider the case of Robert F. Kennedy Jr. and autism. For nearly two decades, Kennedy has waged a campaign that treats autism not as a neurological difference but as a preventable catastrophe, and something that can and should be cured. In 2005, he published articles in Rolling Stone and Salon claiming vaccines caused autism, assertions the scientific community roundly rejected (Oransky 2011). Nearly twenty years later, in April 2025, Kennedy escalated his rhetoric, declaring autism a rapidly growing "epidemic" and dismissing as "indefensible" the well-documented scientific view that rising diagnoses reflect improved recognition, expanded diagnostic criteria, and greater clinical awareness rather than increased prevalence. Insisting "the epidemic is real," and he vowed to identify the specific "environmental toxin" responsible, reframing autism as "a preventable disease."

Kennedy's alternative reality grew more elaborate. Genetic research, he announced, was "a dead end" because "genes do not cause epidemics." Instead, he implicated mould, pesticides, medications, and even ultrasounds as potential culprits. As the newly positioned head of HHS, he promised definitive studies would identify precisely which environmental toxins cause autism, with answers forthcoming by September 2025. It was a bold claim, one that presumed not only certainty about causation but certainty that such causation would reveal itself on a political timeline.

September arrived with predictable results. On the 22nd, Kennedy still maintained that autism constitutes an epidemic, a characterization epidemiologists continue to reject, but the promised identification of specific environmental causes had not materialized. Instead, he backpedalled, tentatively suggesting acetaminophen as a possible factor, despite recent research finding no causal mechanism linking the two, and a new study finding that prenatal aspirin exposure might be associated with a slightly reduced autism risk (Gerlach 2025; Czopek 2025; Ahlqvist et al. 2024). The image he had constructed required constant maintenance, new villains and monsters to sustain the narrative when previous ones failed to materialize.

This strikes painfully close to home. I have brilliant friends and colleagues on the autism spectrum, a remarkable nephew with autism, and heck, I even dated some people with it. Like many parents, autistic individuals, intellectuals, and advocates who have challenged Kennedy's rhetoric, I recognize these people not as victims of an environmental catastrophe but as individuals with distinctive, different, and often extraordinary minds. I feel privileged to know them. Yet my objection extends beyond personal offence. The scientific and epidemiological community's rejection of Kennedy's "epidemic" framing is not merely technical pedantry or “wokeness”, whatever that actually means anymore. When we discuss why scientists reject Kennedy's claims, we find not just errors of fact but something more troubling: the construction of an ideological image that might persist despite contradicting empirical reality, maintained through constant narrative adjustment even as promised evidence fails to materialize. This pattern, in which the creation of images becomes so thorough that the distinction between fabrication and fact becomes increasingly difficult to uphold, echoes some of the concerns Arendt identified in her theory and in relation to the Pentagon Papers.

Category Mistake

Unlike measles, where you either have it or you don't, autism exists on a spectrum: a continuum of traits that shade into typical human variation. Diagnosis relies on behavioural and social observations, typically using standardized assessments such as the ADOS-2 (Autism Diagnostic Observation Schedule) or ADI-R (Autism Diagnostic Interview-Revised), rather than biological tests. The boundary between "autistic" and "not autistic" may rest on social and clinical judgment, making it inherently blurry. The traits in question typically involve differences in social communication, sensory processing, and patterns of behaviour and interests. Because these traits exist on a continuum, any diagnostic cutoff is somewhat arbitrary, but still real. There is no blood test, brain scan, or genetic marker that definitively identifies someone as autistic.

This is why some scientists and researchers refer to a "broad autism phenotype," which describes mild autistic traits found in people who don't meet diagnostic criteria, particularly among family members of autistic individuals (Gerdts and Bernier 2011). Many brilliant thinkers exhibited these traits and might well have received an autism diagnosis, even though some didn’t have the framework available to them at the time, including Alan Turing, Kurt Gödel, Ludwig Wittgenstein, Saul Kripke, and, according to biographical accounts, Immanuel Kant. Several other people might have had it, including Michelangelo, Wolfgang Amadeus Mozart, Nikola Tesla, and Isaac Newton. What's interesting is that many of these individuals' cognitive styles may have directly contributed to their groundbreaking work. For example, when one of my supervisors suggested that Wittgenstein probably had it, it opened the door for me to better understand him, his puzzlement, and his philosophy of communication and language.

Autism is more accurately understood as a neurodevelopmental variation or difference, and in many cases, a valuable one. It is not an infectious or acquired disease that spreads through populations. Kennedy's framing and structure represent a category error.

Neuroscience Might Claim it is a Difference

A noteworthy advancement in autism neuroscience over the past three years is not characterized by a solitary discovery but rather by a paradigm shift. Leading researchers now explicitly oppose the framing of autism as a "disorder" or "disease," supported by neuroimaging evidence indicating that autism involves distinct, rather than deficient, neural architecture. Additionally, genetic studies reveal evolutionary advantages, and findings demonstrate specialized cognitive abilities (Sonuga-Barke 2023). So much so that Simon Baron-Cohen, one of autism research's  influential figures and recipient of the UK's medical research honour, declared in January 2024 on Twitter that autism is "a difference, an identity & a disability, but not a disorder or disease."

Baron-Cohen's 2023 research emphasizes that autism genetics includes genes associated not only with disability but also with talent in pattern recognition and understanding system mechanics. His 2017 editorial in Journal of Child Psychology and Psychiatry asked whether autism should continue being called "disease" or "disorder," arguing evidence at genetic, neural, behavioural, and cognitive levels reveals "both differences, and signs of disability, but not disorder." He distinguished a disability, which requires support, acceptance, and reasonable accommodation, from a disorder, which implies a need for a cure (Baron-Cohen 2017). I might slightly disagree with how he thinks about disability, but his frameworks demand radically different societal responses that perceive it as a “disorder,” whatever that might mean.

In addition, an October 2024 study in Autism Research examined over 11,000 children and found that autism is associated with lower neuron density in cerebral cortex regions involved in memory and reasoning, but also increased neuron density in the amygdala (ScienceDaily 2004; Weber et al. 2024a; Newsweek 2024). This bidirectional pattern, in which some regions have more neurons and others have fewer, might contradict simple deficit models. Lead researcher Dr. Zachary Christensen emphasized these findings provide "a new set of measurements that have shown unique promise in characterizing individuals with autism," framing differences as alternative cellular organization rather than pathology. 

A sophisticated connectivity research analyzed 211 individuals using geodesic distance mapping combined with functional connectomics. Results showed contracted functional connectivity profiles in autism—reduced long-range connections compensated by increased short-range connectivity. Peak differences occurred in the prefrontal, temporal, and parietal cortices, with an effect size of Cohen's d = 0.600 (Weber et al. 2024b).  These patterns positively correlated with IQ scores and remained stable across age groups from childhood to adulthood, suggesting neither developmental delay nor a broken brain but rather a stable alternative architecture organized differently.

Machine learning approaches achieve remarkable precision in identifying neural signatures that distinguish autistic from non-autistic brain organization. A comprehensive systematic review of 146 studies documented classification accuracies of 85-100% using neural signatures, with the highest-performing algorithms achieving 99.91% sensitivity (Gkintoni et al. 2025). This level of precision suggests that autism entails a distinct and consistent neural organization. If neurological degradation were present, the resulting patterns might be inconsistent or unclassifiable. The brain is building something different, not necessarily deficient.

What goes into an "epidemic"?

An epidemic requires, according to the Columbia University Mailman School of Public Health, "an increase, often sudden, in the number of cases of a disease above what is normally expected," typically characterized by rapid onset and identifiable causal agents (Columbia University Mailman School of Public Health 2021).

Several features distinguish genuine epidemics from other patterns of increased diagnosis. First, epidemics exhibit a sudden onset or a sharp rise in cases over a compressed timeframe. Second, they depend on stable measurement systems with consistent diagnostic criteria and detection methods, allowing researchers to distinguish genuine increases from artifacts of changing observation. Third, epidemics reflect actual increases in incidence, meaning more people are developing the condition rather than being recognized more frequently. Fourth, they involve identifiable causes, whether infectious agents or specific environmental exposures. Finally, epidemics display geographic and temporal clustering, with cases concentrated around exposure sources or time periods.

Autism meets none of these criteria.

Sudden onset, which involves a rapid increase in cases over a short time period.

Genuine epidemics involve rapid increases in cases over compressed timeframes. Autism shows no such pattern. Instead, diagnoses have risen gradually over decades, increases that correspond directly to expanded diagnostic criteria, improved clinical recognition, and heightened public awareness (Stein 2025; Smith-Schoenwalder and Johnson 2025; Jacobson 2025). Unlike influenza outbreaks or the COVID-19 pandemic, which produce unmistakable population-wide spikes within weeks or months, autism's apparent "rise" tracks closely with diagnostic and institutional changes rather than any underlying shift in disease prevalence.

A stable measurement, so that you get a consistent diagnostic criterion and detection methods, which will lead to measuring whether it is, in fact, a rapid increase

The epidemic model presupposes stable diagnostic criteria and detection methods, enabling researchers to distinguish genuine incidence changes from measurement artifacts. Autism fails this test entirely. Both diagnostic requirements and detection practices have undergone radical transformations, rendering historical comparisons scientifically invalid.

The evolution of DSM criteria tells the story. Before 1980, autism was barely recognized as a distinct condition. In 1987, diagnostic criteria expanded significantly (Newschaffer et al. 2007). The 1994 revision introduced the spectrum concept, adding Asperger's syndrome and PDD-NOS as separate diagnoses (Newschaffer et al. 2007). By 2013, these categories collapsed into a single ASD diagnosis (Newschaffer et al. 2007). Each revision captured individuals who would have been excluded under previous frameworks.

The impact of these changes is not speculative. Research examining autism cases in Denmark between 1980 and 1991 found that 60% of the apparent increase was due solely to changes in diagnostic criteria, not to any rise in actual prevalence (Smith-Schoenwalder and Johnson 2025; Hansen et al. 2015).

True incidence increase, wherein more people are actually developing the condition, not just more recognition

In genuine epidemics, rising case counts reflect actual increases in disease occurrence. With autism, the evidence overwhelmingly suggests improved identification of existing cases rather than new incidence. Three distinct patterns support this conclusion.

First, diagnostic substitution has played a significant role. As autism diagnoses have increased, intellectual disability diagnoses have decreased proportionally, revealing an inverse relationship. Children who would previously have been categorized under intellectual disability are now more accurately identified as autistic, reflecting a refined understanding rather than a new pathology (Shattuck 2006).

Second, increases concentration precisely where historical underdiagnosis was most severe: among girls (Hull 2017; Cain 2025; D’Mello et al. 2022), racial and ethnic minorities (Mandell et al. 2009; Furfaro 2017; Pham et al. 2022), and high-functioning individuals (Wolff et al. 2022). Earlier, narrower diagnostic frameworks systematically overlooked these groups. Their belated recognition represents the correction of past failures, not an emerging crisis.

Third, the proportion of autistic individuals with co-occurring intellectual disability has remained stable at approximately 40%. If autism were truly becoming more prevalent, we would expect proportional increases across severity levels (CHOP Research Institute 2020; Braun et al. 2015; Nevison 2025). Instead, this stability suggests we are expanding the diagnostic boundaries to include previously missed cases rather than witnessing more severe presentations.

The pattern seems slightly unambiguous to me: better detection of existing variation, not an epidemic of new cases.

Identifiable cause, involving infectious agents or specific environmental exposures.

Autism displays patterns inconsistent with epidemic disease, showing strong genetic loading (80-83% of autism spectrum disorder heritability), and an increased concentration in previously underdiagnosed populations rather than environmental exposure patterns (Sandin et al. 2017; Bai et al. 2019).

True epidemics typically have identifiable triggers, such as contaminated water, infectious agents, or specific toxins. Not surprisingly, despite decades of searching, no single environmental cause explains the increase in autism.

Geographic/temporal clustering, where cases cluster around exposure sources or time periods

The geographic variation is dramatic—California's rate is five times higher than Texas (1 in 19 vs. 1 in 103), reflecting differences in screening programs, diagnostic practices, and service availability rather than underlying population differences.

Actual epidemics usually show consistent increases or cluster around exposure sources. Autism's variation by state, country, and healthcare system reveals it's about who's looking and how, not where the disease is spreading.

What is actually going on here?

Kennedy is engaged in motivated reasoning, driven by decades-long convictions about vaccines and environmental toxins. Rather than examining evidence with genuine openness, he begins with his conclusion and searches selectively for confirmation. This is a classic confirmation bias. His promise to identify "precisely what the environmental toxins are" before conducting any research reveals that his script is already written. The rushed timeline, with predictions of answers within months, suggests the goal is to produce supportive findings rather than to discover the truth.

At the heart of Kennedy's error lies a fundamental category mistake. He treats autism as an acquired disease spreading through populations when the evidence reveals it as a spectrum of neurodevelopmental traits present from birth with strong genetic architecture. His comparison of diagnosis rates across radically different diagnostic eras might resemble comparing thermometers calibrated to different scales and then declaring a "fever epidemic." Multiple studies demonstrate that diagnostic changes, improved awareness, and diagnostic substitution account for substantial portions of observed increases. In some cases, these factors explain the majority. Kennedy dismisses this evidence not because it lacks merit, but because accepting it would undermine the narrative on which he has built his political identity.

This represents what scholars call policy-based evidence-making: deciding on action first, then selectively gathering evidence to justify it afterward. Kennedy has constructed a public persona around exposing "uncomfortable truths" that big establishments allegedly hide. Acknowledging that autism increases reflect better diagnosis rather than environmental catastrophe would require abandoning this entire framework. The admission would unravel his years of advocacy and undermine his credibility with his supporters, who have trusted his narrative about hidden dangers and institutional cover-ups.

The consequences extend beyond rhetorical missteps. As the Autism Society warns, Kennedy's approach "wastes valuable resources," "perpetuates stigma," and substitutes "speculation and oversimplified timelines" for rigorous science. Most damagingly, he harms the very population he claims to help, by framing autism as a preventable tragedy and something to cure, rather than human neurodiversity requiring social support and accommodation, Kennedy reinforces the stigma that autistic people, their allies, advocates and their families work to overcome.

The real problem here is not autism. It is the refusal to accept evidence that contradicts his prior beliefs, and the willingness to sacrifice scientific integrity and human dignity in service of a predetermined narrative. His predetermined narrative. His image-making.

Works Cited

Ahlqvist, Viktor H., Hugo Sjöqvist, Christina Dalman, et al. 2024. “Acetaminophen Use During Pregnancy and Children’s Risk of Autism, ADHD, and Intellectual Disability.” JAMA 331 (14): 1205–14. https://doi.org/10.1001/jama.2024.3172.

Bai, Dan, Benjamin Hon Kei Yip, Gayle C. Windham, et al. 2019. “Association of Genetic and Environmental Factors With Autism in a 5-Country Cohort.” JAMA Psychiatry 76 (10): 1035–43. https://doi.org/10.1001/jamapsychiatry.2019.1411.

Baron-Cohen, Simon. 2017. “Editorial Perspective: Neurodiversity - a Revolutionary Concept for Autism and Psychiatry.” Journal of Child Psychology and Psychiatry, and Allied Disciplines 58 (6): 744–47. https://doi.org/10.1111/jcpp.12703.

Braun, Kim Van Naarden, Deborah Christensen, Nancy Doernberg, et al. 2015. “Trends in the Prevalence of Autism Spectrum Disorder, Cerebral Palsy, Hearing Loss, Intellectual Disability, and Vision Impairment, Metropolitan Atlanta, 1991–2010.” PLOS ONE 10 (4): e0124120. https://doi.org/10.1371/journal.pone.0124120.

Cain, Rachel. 2025. “New Study Reveals Gaps in Autism Diagnosis for Girls.” Text. Medical School, August 29. https://med.umn.edu/news/new-study-reveals-gaps-autism-diagnosis-girls.

CHOP Research Institute. 2020. “Intellectual Disability and ASD.” June 15. https://www.research.chop.edu/car-autism-roadmap/intellectual-disability-and-asd.

Columbia University Mailman School of Public Health. 2021. “Epidemic, Endemic, Pandemic: What Are the Differences?” February 19. https://www.publichealth.columbia.edu/news/epidemic-endemic-pandemic-what-are-differences.

Czopek, Madison. 2025. “Research Doesn’t Show Using Tylenol during Pregnancy Causes Autism. Here Are 5 Things to Know.” PBS News, September 17. https://www.pbs.org/newshour/health/research-doesnt-show-using-tylenol-during-pregnancy-causes-autism-here-are-5-things-to-know.

D’Mello, Anila M., Isabelle R. Frosch, Cindy E. Li, Annie L. Cardinaux, and John D.E. Gabrieli. 2022. “Exclusion of Females in Autism Research: Empirical Evidence for a ‘Leaky’ Recruitment‐to‐research Pipeline.” Autism Research 15 (10): 1929–40. https://doi.org/10.1002/aur.2795.

Furfaro, Hannah. 2017. “Race, Class Contribute to Disparities in Autism Diagnoses.” The Transmitter: Neuroscience News and Perspectives, November 20. https://www.thetransmitter.org/spectrum/race-class-contribute-disparities-autism-diagnoses/.

Gerdts, Jennifer, and Raphael Bernier. 2011. “The Broader Autism Phenotype and Its Implications on the Etiology and Treatment of Autism Spectrum Disorders.” Autism Research and Treatment 2011: 545901. https://doi.org/10.1155/2011/545901.

Gerlach, Alexandra. 2025. “HHS Report Expected to Link Prenatal Tylenol Use and Autism Risk | Pharmacy Times.” November 5. https://www.pharmacytimes.com/view/hhs-report-expected-to-link-prenatal-tylenol-use-and-autism-risk.

Gkintoni, Evgenia, Maria Panagioti, Stephanos P. Vassilopoulos, Georgios Nikolaou, Basilis Boutsinas, and Apostolos Vantarakis. 2025. “Leveraging AI-Driven Neuroimaging Biomarkers for Early Detection and Social Function Prediction in Autism Spectrum Disorders: A Systematic Review.” Healthcare 13 (15): 1776. https://doi.org/10.3390/healthcare13151776.

Hansen, Stefan N., Diana E. Schendel, and Erik T. Parner. 2015. “Explaining the Increase in the Prevalence of Autism Spectrum Disorders: The Proportion Attributable to Changes in Reporting Practices.” JAMA Pediatrics 169 (1): 56–62. https://doi.org/10.1001/jamapediatrics.2014.1893.

Hull, Laura. 2017. “The Gender Gap in Autism.” Ambitious about Autism, September 27. https://www.ambitiousaboutautism.org.uk/about-us/media-centre/blog/girls-and-autism/gender-gap-autism.

Jacobson, Louis. 2025. “Fact-Checking Robert F. Kennedy Jr.’s Statements on Autism.” PBS News, April 23. https://www.pbs.org/newshour/politics/fact-checking-robert-f-kennedy-jr-s-statements-on-autism.

Mandell, David S., Lisa D. Wiggins, Laura Arnstein Carpenter, et al. 2009. “Racial/Ethnic Disparities in the Identification of Children With Autism Spectrum Disorders.” American Journal of Public Health 99 (3): 493–98. https://doi.org/10.2105/AJPH.2007.131243.

Nevison, Cynthia. 2025. Time Trends in United States Autism Prevalence with Co-Occurring Intellectual Disability: Is There a Signature of Thimerosal? - Science, Public Health Policy and the Law. Public Health. June 25. https://publichealthpolicyjournal.com/time-trends-in-united-states-autism-prevalence-with-co-occurring-intellectual-disability-is-there-a-signature-of-thimerosal/.

Newschaffer, Craig J., Lisa A. Croen, Julie Daniels, et al. 2007. “The Epidemiology of Autism Spectrum Disorders.” Annual Review of Public Health 28 (1): 235–58. https://doi.org/10.1146/annurev.publhealth.28.021406.144007.

Newsweek. 2024. “Neurons Different in Children with Autism, Study Finds.” October 10. https://www.newsweek.com/neurons-different-children-autism-study-1967219.

Oransky, Ivan. 2011. “Salon Retracts 2005 Robert F. Kennedy Jr. Piece on Alleged Autism-Vaccine Link.” Retraction Watch, January 16. https://retractionwatch.com/2011/01/16/salon-retracts-2005-robert-f-kennedy-jr-piece-on-alleged-autism-vaccine-link/.

Pham, Hoangmai H., Neil Sandberg, Jeff Trinkl, and Johnston Thayer. 2022. “Racial and Ethnic Differences in Rates and Age of Diagnosis of Autism Spectrum Disorder.” JAMA Network Open 5 (10): e2239604. https://doi.org/10.1001/jamanetworkopen.2022.39604.

Sandin, Sven, Paul Lichtenstein, Ralf Kuja-Halkola, Christina Hultman, Henrik Larsson, and Abraham Reichenberg. 2017. “The Heritability of Autism Spectrum Disorder.” JAMA 318 (12): 1182–84. https://doi.org/10.1001/jama.2017.12141.

ScienceDaily. 2004. “Neurons Look Different in Children with Autism, Research Finds.” https://www.sciencedaily.com/releases/2024/10/241009121634.htm.

Shattuck, Paul T. 2006. “The Contribution of Diagnostic Substitution to the Growing Administrative Prevalence of Autism in US Special Education.” Pediatrics 117 (4): 1028–37. https://doi.org/10.1542/peds.2005-1516.

Smith-Schoenwalder, Cecelia, and Steven Ross Johnson. 2025. “4 Takeaways From RFK Jr.’s Press Conference on CDC Autism Report.” US News & World Report, April 16. //www.usnews.com/news/health-news/articles/2025-04-16/4-takeaways-from-rfk-jr-s-press-conference-on-cdc-autism-report.

Sonuga-Barke, Edmund J. S. 2023. “Paradigm ‘Flipping’ to Reinvigorate Translational Science: Outlining a Neurodevelopmental Science Framework from a ‘Neurodiversity’ Perspective.” Journal of Child Psychology and Psychiatry 64 (10): 1405–8. https://doi.org/10.1111/jcpp.13886.

Stein, Rob. 2025. “RFK Jr. Calls Autism an ‘epidemic’ and Launches Effort to Find ‘Environmental’ Cause.” Children’s Health. NPR, April 17. https://www.npr.org/2025/04/17/nx-s1-5366846/rfk-jr-calls-autism-an-epidemic-and-launches-effort-to-find-environmental-cause.

Weber, Clara F., Valeria Kebets, Oualid Benkarim, et al. 2024a. “Contracted Functional Connectivity Profiles in Autism.” Molecular Autism 15 (1): 38. https://doi.org/10.1186/s13229-024-00616-2.

Weber, Clara F., Valeria Kebets, Oualid Benkarim, et al. 2024b. “Contracted Functional Connectivity Profiles in Autism.” Molecular Autism 15 (1): 38. https://doi.org/10.1186/s13229-024-00616-2.

Wolff, Nicole, Sanna Stroth, Inge Kamp-Becker, Stefan Roepke, and Veit Roessner. 2022. “Autism Spectrum Disorder and IQ – A Complex Interplay.” Frontiers in Psychiatry 13 (April). https://doi.org/10.3389/fpsyt.2022.856084.