“Masking Is a Myth, Late Diagnoses Are Wrong, and Non-Speaking Autistic People May Be Lying” - What Uta Frith Got Wrong
Professor Uta Frith has spent more than fifty years studying autism. She was among the first researchers to identify the cognitive profile underlying autistic experience, her work shaped a generation of clinical understanding, and her influence on how autism came to be recognised in the UK is, by any reasonable measure, substantial. None of that is in dispute. What is in dispute is what she said in a recent TES interview and in a separate podcast conversation with clinical psychologists Dr Naomi Fisher and Dr Danielle Drinkwater for their new ‘Let’s Talk Neurosense’ series - and what those claims, coming from her specifically, are likely to do.
Across both, Frith argued that the autism spectrum has collapsed through overextension, that the late diagnosis of women and girls probably reflects misidentification rather than previously missed autism, that the concept of masking has no scientific basis and depends on unfalsifiable self-report, and that non-speaking autistic people who produced written accounts of their inner lives were, in her view (she acknowledged she had no proof) - likely engaged in conscious or subconscious fraud. The diagnostic credibility of autism, she concluded, had been severely damaged, if not destroyed, by the cultural adoption of these ideas.
It is worth noting that in the podcast, Frith was not challenged on any of this. Not on masking, not on sensory processing, not on the claims about non-speaking autistic people. For a conversation presented as a forum for nuanced debate about neurodiversity, that absence is notable - several people who had listened to it said as much directly in the comments. What the listener gets instead is a sympathetic platform for a set of claims that deserved considerably more scrutiny than they received.
These claims are not peripheral observations from someone out of the field. They are the considered positions of the person most institutionally authorised to make them. They will travel, and they will arrive, stripped of their caveats, in classrooms and clinical assessments and local authority decision-making, as permission to dismiss what autistic people say about their own experience - even more than there already is.
The question worth examining is not whether Frith has legitimate concerns about diagnostic boundary questions - she does, and so do many researchers working in this area. The question is whether the specific argument she has built from those concerns holds up, and whether the way she has chosen to handle the evidence is as rigorous as the authority she carries would lead you to expect.
On masking
Frith’s argument about masking runs as follows: the concept depends entirely on subjective self-report; there is no way to falsify a claim that someone is masking; therefore it has no scientific basis and its widespread acceptance by clinicians and researchers has been dangerous. The exhaustion people attribute to masking, she suggests, could have many other explanations - chronic fatigue, anxiety, functional neurological disorder - and the concept has functioned as a cultural meme, drawing people toward an autism diagnosis as an explanation for distress that may have other causes.
The methodological concern buried in this argument is a real one, and it applies much more widely than Frith applies it. The unfalsifiability objection - that you cannot tell someone they are not experiencing what they report experiencing - is not unique to masking. It applies equally to pain scales, depression inventories, anxiety assessments, and the majority of psychiatric diagnostic instruments, almost all of which depend substantially on self-report and cannot be independently verified by external observation. Frith applies it to masking. She also applies it to sensory processing - describing sensory issues as a “big deal coming to the foreground” among “new autism seekers,” with the implication that this too is a culturally absorbed symptom rather than a documented feature of autistic neurology. If this standard were applied consistently across clinical psychology, the diagnostic manual would be considerably thinner than it is. Frith does not apply it consistently.
The more significant problem with her position is that it is empirically wrong. The concept of autistic masking had been named and described by autistic people themselves considerably earlier, but it became a sustained focus of academic research in the 2010s - a body of work that was, in significant part, prompted by what autistic communities had already been documenting about their own experience. Frith’s claim that masking has “no scientific basis” is therefore both factually incorrect and historically compressed: it erases the decade of autistic community knowledge that preceded and in many ways generated the academic interest she dismisses. Masking in autistic people - defined as the conscious or unconscious suppression of autistic characteristics, camouflaging of difficulties, and social mimicry used to fit into neurotypical environments - has a peer-reviewed research literature built on that foundation. Cage and Troxell-Whitman (2019), published in the Journal of Autism and Developmental Disorders, examined the reasons, contexts and costs of camouflaging in autistic adults, identifying distinct motivational patterns and documenting significant associations between higher camouflaging and poorer mental health outcomes. Hull and colleagues have published multiple studies developing and validating the Camouflaging Autistic Traits Questionnaire, producing a measurement instrument specifically designed to operationalise the construct reliably enough to study it. The Royal College of Psychiatrists, in its 2020 guidance on the psychiatric management of autism in adults, defined autistic burnout explicitly as a state of exhaustion resulting from coping with social interaction including masking, and linked it to heightened risk of suicide. Frith’s call for “hard research” on masking is not a call for something that doesn’t exist, it is a call, issued as though the literature doesn’t exist, for more of what has already been done.
There is a further problem with how Frith approaches clinical assessment that the masking literature makes visible. In the podcast, she identifies certain features -smooth conversation, humour, irony, the ability to read between the lines - as indicators that may weigh against an autism diagnosis. The difficulty is that these are precisely the capacities that decades of effortful masking can produce. An autistic person who has spent years learning to perform smooth conversation, to deploy humour strategically, to read social cues through consciously applied rules rather than instinctive processing, will present to a clinician as someone who does not look autistic. That is not evidence they are not autistic, it is evidence the masking worked. Using the observable outputs of successful masking as clinical evidence against the diagnosis is not rigorous assessment, but the circular confirmation problem Frith warns against - applied, without apparent awareness, to her own preferred framework.
Frith also says, in the podcast, that “I expect we could say we are all masking, all the time, trying to adapt to our society’s norms” - that social adaptation is universal and there is nothing distinctively autistic about the phenomenon. This is the version of the argument most likely to feel persuasive to people who haven’t encountered the research, because it contains something true: all humans modulate their behaviour in social contexts. What it misses is that autistic masking is not social modulation. It is the suppression of the actual way one’s nervous system functions - stimming that regulates sensory overload, communication styles that are genuinely different rather than merely informal, processing needs that require time or stillness that the environment doesn’t permit - in order to avoid the social consequences of being visibly autistic. This distinction is important because the cost is categorically different. The research on autistic burnout - Arnold et al. (2023), Raymaker et al. (2020), Siggers and Day (2024) - documents what sustained masking produces: not mild tiredness after a long day, but functional collapse, loss of previously held skills, withdrawal from life, and in some cases suicidal crisis. These are measurable outcomes with documented patterns. Frith’s suggestion that this exhaustion probably has other explanations is speculation - her word - offered without evidence, against a body of peer-reviewed literature she dismisses without engaging with it.
There is also a logical problem with the masking argument that sits alongside the empirical one. Frith’s broader thesis is that autism has become a desirable cultural identity - a meme that people have absorbed and sought out as an explanation for ordinary distress. But if that is true, why are so many of the same people hiding visible evidence of that identity at enormous personal cost, in every social context, until they collapse from the effort? Someone pursuing an autism diagnosis purely as a cultural identity - which is Frith's own characterisation of what is happening - has no reason to be simultaneously, compulsively, exhaustingly suppressing every outward sign of that same identity. The direction of the incentive runs the wrong way. You cannot coherently argue both that autism is a label people are culturally drawn towards, and that the mechanism by which they concealed their autism for decades has no basis in reality.
On late diagnosis and women
The argument that late-diagnosed women are probably misidentified rests on the claim that if someone were genuinely autistic, something would have been noticed earlier - by professionals, by schools, by the diagnostic systems that existed when they were children. This argument has a structural problem that Frith does not address, and it is not a subtle one.
The diagnostic tools and clinical criteria used throughout the latter decades of the twentieth century were developed predominantly from research on boys. The male-to-female ratio in early autism research was approximately 4:1, and the behavioural profile that became the basis for diagnosis reflected that sample. Frith herself acknowledges, in the podcast, that the ratio has shifted - closer to 3:1 - but draws from this the conclusion that recognition of autism in girls has always existed, at least for early-diagnosed cases. What she does not address is the implication of that shift for the generations of women who were not early-diagnosed, who passed through those diagnostic systems without being found, and who are now presenting in adulthood. Girls and women whose autistic characteristics manifested differently - through social mimicry, through the kind of exhausting rule-learning that looks like compliance, through internalising distress rather than externalising it - were systematically not captured by instruments that were not built to find them. Loomes, Hull and Mandy’s 2017 systematic review and meta-analysis, published in the Journal of the American Academy of Child and Adolescent Psychiatry, documented this diagnostic bias directly. The reason nothing was noticed is not that there was nothing to notice. The reason nothing was noticed is that the people doing the noticing were looking for something different, in a profile that girls were much less likely to display.
Frith’s response to the increase in late-diagnosed women is to suggest that this population probably doesn’t have autism, and probably has anxiety or chronic fatigue or something else that the autism meme has absorbed. What she does not engage with is the converse possibility - that women who were dismissed as anxious or sensitive or simply socially awkward for thirty or forty years, and who now receive a late diagnosis that finally makes sense of what they have been experiencing, are not victims of diagnostic fashion, they are the people her generation of specialists missed. Ann Memmott, a research consultant who is herself autistic, put this directly in response to the TES article: the increase in late-diagnosed adults is not frightening, it is what finding the people who were always there looks like.
It is also worth being precise about what diagnosis actually does in this context, because Frith repeatedly conflates two things that are not the same. Pathologising is what she is doing - framing an increasing number of people as having absorbed a cultural contagion, as identity-seeking, as a problem to be explained away. Diagnosis is the opposite: it is the recognition that what someone has been living with has a name, a documented pattern, and implications for what support they actually need. The people seeking late diagnoses are not seeking pathology, they are seeking, often after decades of confusion and self-blame, an accurate description of how their nervous system works.
On what counts as evidence
Running beneath both of these arguments is a third one, and it is the one that does the most damage in practice. Frith repeatedly dismisses self-report as insufficiently rigorous - she describes lived experience as a “buzz word,” questions whether it can constitute objective evidence, and contrasts it unfavourably with what she calls “proper hard-line research.” In the podcast, she expresses scepticism about whether people can reliably reflect on their own inner states at all, arguing that what people say about themselves “is not a pure, one-source authentic lived experience” but rather “a construct from the whole of society.”
When pressed on what rigorous research actually looks like, she specifies: falsifiable hypotheses, strictly controlled comparison groups, hard biomarker data. Research that does not meet this standard - which includes the majority of the existing masking and burnout literature, the qualitative studies documenting autistic women’s diagnostic experiences, and the lived experience accounts that inform clinical guidance - she regards as not vigorous enough to count. The compassion motivating some of this work, she says, is understandable but dangerous, because compassion without rigour produces circular confirmation.
The problem with this as a response to a body of peer-reviewed, published, replicated research is that it isn’t an engagement with the evidence. It is a pre-emptive reclassification of the people who produced it as epistemologically compromised. The Royal College of Psychiatrists absorbed the meme. Hull and colleagues absorbed the meme. The researchers documenting autistic burnout absorbed the meme. The clinicians whose guidance now explicitly names masking as a mechanism are operating from compassion rather than science. None of their work therefore counts. This is a position that is, in her own term, unfalsifiable: no evidence from within the current field can challenge it, because the field itself has been designated as contaminated. She has, in other words, applied to the entire research literature precisely the logical move she uses to dismiss masking - and apparently without noticing the irony.
This position, applied to autistic people specifically, produces a closed loop. Autistic people’s accounts of their own experience are inadmissible because they are subjective. The observable behaviours that clinicians can measure don’t capture the full picture, but the picture that isn’t captured can’t be cited because it’s self-reported. The diagnostic tools built on a biased sample don’t find the women and girls who weren’t in that sample, but the fact that those women and girls report experiences consistent with autism can’t count as evidence because self-report is unreliable. And the non-speaking autistic people who produce written accounts of their inner lives - whose communication represents, for many of them, the primary mechanism by which they can be understood at all - are, in Frith’s framing, probably not to be believed. It is worth being exact about what she actually argues here, because the claim is often softened in summary. She invokes the facilitated communication controversy - a legitimate and documented concern about a specific historical practice in which a facilitator’s movements could contaminate or replace a non-speaking person’s communication - and extends it into a generalised suspicion of non-speaking autistic people’s independent written accounts. These are not the same thing, and the conflation does real harm to people whose testimony was hard-won and is not the product of facilitation.
What this framework cannot accommodate is the possibility that the forms of evidence it excludes are the forms of evidence most necessary to understand the people it has historically failed to understand. Autistic experience is, by definition, internally located in significant ways - the sensory processing differences that Frith also dismissed as a culturally absorbed complaint, the social exhaustion, the way that environments which look manageable from the outside produce cumulative internal cost that only becomes visible in collapse. Demanding that this experience justify itself through methods designed to detect externally observable behaviour is not rigour, it is asking the wrong question and then counting the absence of an answer as proof.
There is also a deeper disagreement underneath all of this, and it is worth naming directly because it shapes the entire argument. Frith is defending a classical medical-model use of diagnosis: narrow category, clear external impairment, scepticism about self-report, hard observable data. Much of the current neurodiversity-informed clinical practice uses diagnosis differently - as an explanatory frame for lived experience, access needs, and the kind of understanding that changes how someone is treated by the systems around them. These frameworks are genuinely in tension, and the disagreement between them is not only about data, it is about what diagnoses are for. Frith’s model has real strengths for certain research purposes, but it has significant limitations as a description of how diagnosis actually functions for the people who receive one - particularly for those whose primary difficulties are internal, cumulative, and only visible to others when the effort of concealment finally fails.
On the standard she applies to others
It is worth pausing on what Frith’s demand for rigorous research actually looks like when applied to her own foundational work, because she addresses this briefly in the podcast and the implications are more significant than she appears to register.
She describes the early autism research - the work that produced the diagnostic criteria she now treats as the authoritative baseline - as having been conducted on a very small pool of participants, the same children studied repeatedly, in ways she herself acknowledges were methodologically problematic. The early field had everything she now criticises the current literature for lacking: no controlled comparison groups, no biomarkers, no population-level sampling, heavy reliance on clinical observation of a narrow and unrepresentative cohort. The rigorous standard she is holding the masking and lived experience research to is a standard her own foundational field could not have met. She seems to regard this as interesting historical context rather than a problem for her argument. It is, in fact, a significant one.
The biomarker argument contains a similar tension. Frith argues, at some length, that the absence of clear biological markers for autism is evidence that the diagnostic category has been diluted - that a more coherent, appropriately bounded diagnosis would produce cleaner neurological correlates. Two things are worth saying about this. The first is that the absence of a biomarker does not, as a general principle of clinical medicine, undermine diagnostic validity. Many well-established and clinically meaningful conditions - including migraine, fibromyalgia, and functional neurological disorder - have no single confirming biomarker. The argument “we cannot find a biomarker, therefore the diagnosis may not be real” is not one Frith would likely apply to those conditions, and there is no principled reason to apply it selectively to autism. The second is that she herself acknowledges, within the same discussion, that the brain is extraordinarily complex and that this complexity may itself explain why biomarkers have proved so elusive. These two positions cannot both be doing the work she wants them to do simultaneously. Either the absence of biomarkers tells us the category is too broad, or it tells us the brain is more complex than our instruments can currently capture. Frith uses both explanations in the same conversation without appearing to notice that they contradict each other - which is, for someone who has just spent considerable time demanding that others falsify their hypotheses and acknowledge when they might be wrong, a notable oversight.
On the world we are actually living in
Frith’s explanation for rising autism diagnoses centres on destigmatisation, the influence of film and television, the internet, and the online peer-finding that allows people to identify with others who share their experiences. She acknowledges these are speculative rather than evidenced. She does not acknowledge that they are also selective.
The world has changed in other ways alongside these, and those changes have not received the same scrutiny in her account. The pace, noise, unpredictability and sensory intensity of contemporary life have increased substantially over the decades in which autism diagnoses have risen. The unstructured time, low-demand environments, and tolerance for different ways of functioning that historically gave neurodivergent people room to manage have narrowed significantly: schools have become more regimented, more focused on measurable output, less able to accommodate children who learn differently or need more time. Open-plan offices, constant connectivity, social media performance, and an economic environment that demands flexibility and rapid adaptation have all placed particular pressure on nervous systems that function best with predictability, clear expectations, and the ability to regulate at their own pace. If autistic neurology is - as the research consistently suggests - particularly poorly matched to environments characterised by sensory overload, unpredictability, and relentless social demand, then the appropriate question is not why more people are seeking diagnosis, but why we would expect anything else.
Frith asks, rhetorically, whether the logical endpoint of expanding diagnostic categories is a situation in which the entire population requires psychiatric or psychological support. She presents this as a reductio ad absurdum. The available mental health data across all age groups in wealthy countries does not obviously support treating it as absurd. The correct response to widespread distress is not to narrow the categories through which it can be recognised and supported, it is to take seriously the question of what is producing it.
There is also a category distinction buried in this argument that her framing obscures. Frith is pathologising - treating rising autism recognition as a symptom of cultural contagion, diagnostic fashion, and lowered thresholds. Diagnosis, properly understood, is the opposite of this. It does not create the distress it describes. It names it, which is the prerequisite for understanding what might actually help.
On the timing
Frith’s interview and podcast were not published in a vacuum. They arrived in early 2026, weeks after the Government’s Schools White Paper proposed replacing legally enforceable Education, Health and Care Plans with Individual Support Plans of considerably less established legal weight – with both explicitly reserved for only the ‘most complex needs.’ They arrived after the Measure What Matters investigation published its review of 25 neurodivergent children whose deaths between 2020 and 2024 were preceded by documented, repeated failures in SEND provision. They arrived as that same White Paper described autism and ADHD - the diagnoses present in every single detailed case study in that investigation - as “predictable” needs that mainstream schools, better resourced, should be able to absorb without statutory plans.
It is worth sitting with what “most complex” is doing in that framing. The children whose families I work with - the ones who cannot sustain a school day without it costing more than they have, whose distress is read as behaviour, whose needs are real and documented and consistently unmet - are, in the vast majority of cases, autistic children whose presentation includes precisely the things that are now being publicly questioned: masking and its consequences, sensory processing difficulties, demand avoidance, late or missed diagnosis. The White Paper wants to reserve statutory protection for a narrower and more visible category of need. And in the same period, a cluster of high-profile voices - Stock on PDA, Frith on masking and late diagnosis - is publicly narrowing what counts as genuine need in the first place. These are not coordinated arguments. But arguments do not need to be coordinated to do convergent work. They need only to point in the same direction, and in early 2026, they do.
Frith likely did not intend her interview as a contribution to a policy argument about whether disabled children deserve legally enforceable rights. She is making what she regards as a scientific case, but the scientific case has consequences in the world it enters, and the world it has entered is one in which the question of what is owed to neurodivergent children is being actively contested, at pace, by people who will find her arguments very useful indeed.
A note on the children this affects
The families I work with are not, in the main, people who have sought out a diagnosis as an identity or an explanation for ordinary social anxiety. They are families whose children could not get through a school day without it costing more than they had to give, who spent years being told that their child was fine - fine in school, coping, managing - while watching something collapse at home each evening that they couldn’t name and couldn’t get anyone to take seriously. They are parents who sat in meetings where their account of what they were seeing was weighed against a school’s observation that the child seemed calm in class, and whose account lost. They are autistic adults, many of them women (myself included,) who spent decades understanding themselves through the framework available to them - anxious, oversensitive, exhausting, not quite right in ways nobody could explain - and who found in a late diagnosis not a fashionable identity but the first coherent description of the way their nervous system had always worked.
Frith’s framework has no satisfactory account of these people. Her argument requires them to be the product of lowered diagnostic thresholds and cultural contagion, their exhaustion probably attributable to something else, their self-reports epistemologically suspect, their late diagnoses a symptom of a field that has lost its way. The research does not support this account. And the children who are currently in school, accruing the kind of unmet need that the Measure What Matters data documents in forensic detail, do not have the luxury of waiting for the field to decide whether their experience counts as evidence.
Frith’s contribution to autism research over fifty years is not in question, however, what is in question is whether the specific claims she made are accurate, whether the evidence she invokes supports the conclusions she draws, and whether the dismissiveness with which she treats the forms of knowledge she finds inconvenient is consistent with the standard she sets for others. On all three counts, the answer is no.
REFERENCES
On masking — peer-reviewed literature
Cage, E. and Troxell-Whitman, Z. (2019). Understanding the reasons, contexts and costs of camouflaging for autistic adults. Journal of Autism and Developmental Disorders, 49(5), 1899–1911.
Hull, L., Mandy, W., Lai, M-C., Baron-Cohen, S., Allison, C., Smith, P. and Petrides, K.V. (2019). Development and validation of the Camouflaging Autistic Traits Questionnaire (CAT-Q). Journal of Autism and Developmental Disorders, 49(3), 819–833.
Bradley, L., Shaw, R., Baron-Cohen, S. and Cassidy, S. (2021). Autistic adults' experiences of camouflaging and its perceived impact on mental health. Autism in Adulthood, 3(4), 338–348.
Scheeren, A.M., Nieuwenhuis, S., Crane, L., Roke, Y. and Begeer, S. (2025). Masking, social context and perceived stress in autistic adults: An ecological momentary assessment study. Autism, 29(12), 3002–3013. https://doi.org/10.1177/13623613251353358
Cook, J.M., Crane, L. and Mandy, W. (2023). Dropping the mask: It takes two. Autism, 1–12. https://doi.org/10.1177/13623613231183059
Evans, J.A., Krumrei-Mancuso, E.J. and Rouse, S.V. (2024). What you are hiding could be hurting you: Autistic masking in relation to mental health, interpersonal trauma, authenticity, and self-esteem. Autism in Adulthood, 6(2), 229. https://doi.org/10.1089/aut.2022.0115
Alaghband-rad, J., Hajikarim-Hamedani, A. and Motamed, M. (2022). Camouflage and masking behavior in adult autism. Frontiers in Psychiatry, 13. https://doi.org/10.3389/fpsyt.2022.1108110
On autistic burnout
Siggers, G. and Day, B. (2024). Beyond school avoidance: Recognising, identifying, and addressing autistic burnout in children. BJPsych Open, 10(S1). https://doi.org/10.1192/bjo.2024.433
Raymaker, D.M., Teo, A.R., Steckler, N.A., Lentz, B., Scharer, M., Delos Santos, A., Kapp, S.K., Hunter, M., Joyce, A. and Nicolaidis, C. (2020). "Having all of your internal resources exhausted beyond measure and being left with no clean-up crew": Defining autistic burnout. Autism in Adulthood, 2(2), 132–143.
Arnold, S.R.C., Higgins, J.M., Weise, J., Desai, A., Pellicano, E. and Trollor, J.N. (2023). Confirming the nature of autistic burnout. Autism, 27(7), 1906–1918.
On late diagnosis and gender bias
Loomes, R., Hull, L. and Mandy, W.P.L. (2017). What is the male-to-female ratio in autism spectrum disorder? A systematic review and meta-analysis. Journal of the American Academy of Child and Adolescent Psychiatry, 56(6), 466–474.
On clinical guidance explicitly naming masking
Royal College of Psychiatrists (2020). Psychiatrists' Support Service: Autism in adults — guidance for psychiatrists. RCPsych.
