When 'reasonable' means neurotypical
A paper published this month in The American Journal of Bioethics exposes a structural flaw in how autism research obtains informed consent. The argument is US-focused, built around federal guidance requiring that consent disclosures meet a “reasonable person standard” — providing whatever information an “average” person would want to know. But the structural problem applies wherever consent frameworks are designed around assumptions that autistic participants do not share.
The “reasonable person” in legal and ethical guidance is not a neutral figure. Historic characterisations include the “man on the Clapham omnibus” and “the man who takes the magazines at home and in the evening pushes the lawn-mower in his shirt sleeves.” The imagined average is upper-middle-class, employed, neurotypical. Their concerns become the baseline. Anyone whose informational needs differ from this hypothetical average gets systematically under-informed.
The Ohio State University research team — which includes autistic co-authors — argues that because the neurodiversity paradigm is widely held in the autistic community, and because it fundamentally shapes what information autistic people need to make participation decisions, consent processes that ignore it fail on their own stated terms. The purpose of informed consent is to enable autonomous choice. If disclosure requirements are indexed to what neurotypical people would want to know, autistic participants cannot make genuinely informed decisions about research that directly concerns them.
UK research ethics operates under different governance — the UK Policy Framework for Health and Social Care Research, NHS Research Ethics Committees — but the structural problem transfers. Consent frameworks here similarly assume a baseline participant whose concerns align with institutional and medical model framing. The question the bioethicists pose applies equally: reasonable to whom?
Three concerns consent forms won't disclose
The paper identifies three categories of concern that autistic people frequently hold about autism research but that consent processes do not address.
First, pathologising language. A systematic review found that over 70% of terms in autism research articles use traditional medical language rather than neuro-affirming alternatives. The DSM-5 definition describes “persistent impairment” and “deficits.” This framing shapes public perception, employer attitudes, and interpersonal relationships. Many autistic people would want to know what language researchers will use in publications before deciding whether to participate. But current guidance requires disclosure of risks to participants, not downstream harms to the broader community. The employment consequences of pathologising language fall outside disclosure requirements.
Second, research aimed at reducing autistic traits. A substantial body of autism research develops interventions designed to make autistic people appear more neurotypical — reducing repetitive behaviours, teaching social rules to enable “passing,” measuring success by how non-autistic someone presents. Many autistic scholars argue that such research reinforces unjust expectations and that masking carries documented mental health risks including burnout and suicidality. But the “reasonable person” — imagined as sharing medical model assumptions — is unlikely to see anything problematic about research helping autistic people act more neurotypically. Such concerns need not be disclosed because the average person would not hold them.
Third, researcher positionality. The paper notes that autistic advocates have raised concerns about researchers whose previous work they view as harmful — including Simon Baron-Cohen, whose empathy-deficit framing shaped decades of autism discourse, and historical figures like Ole Ivar Lovaas, who advocated extreme corporal punishment in early behavioural interventions. The autistic community’s objections to Spectrum 10K centred substantially on Baron-Cohen’s involvement. But under current guidance, researcher views on autism and previous publication records need not be disclosed. The “reasonable person” would not distrust a researcher for holding medical model views because the “reasonable person” likely shares them.
Spectrum 10K proved the gap exists
UK audiences need not look to US bioethics for evidence that consent frameworks fail autistic participants. Spectrum 10K — a Cambridge-based genetic research project paused in 2021 following intense backlash from the autistic community — demonstrated the gap in practice.
The project aimed to collect genetic data from 10,000 autistic people in the UK. Autistic advocates raised concerns about downstream uses of such data, including the possibility of prenatal testing and selective termination. They raised concerns about Baron-Cohen’s leadership given his previous work. They raised concerns about vague aims and broad data-sharing provisions. The consent process had met regulatory requirements. The study had ethics approval. The backlash came anyway — because the concerns autistic people actually held were not the concerns consent frameworks required researchers to address.
The bioethicists frame this precisely. Consent guidance requires disclosure of risks to participants. It does not require disclosure of potential downstream consequences to the autistic community — genetic data being used to seek a “cure,” research reinforcing discriminatory norms, publications shaping public perception in ways that harm autistic employment prospects. These are not covered because the imagined “reasonable person” would not hold such concerns. The neurodiversity paradigm remains a minority viewpoint in the general population. Consent frameworks built around majority assumptions systematically exclude minority concerns.
The paper notes that providing participants the opportunity to ask questions does not resolve this. People often do not know what they do not know. Substantial knowledge of research history and scientific literature would be required for a participant to know what questions to ask about researcher positionality or downstream data uses. The opportunity to ask questions does not substitute for disclosure of information participants need.
Policy as alibi in research ethics
The familiar pattern. A process exists. Requirements are met. Boxes are ticked. Yet the actual needs of the people the process claims to serve go unaddressed.
Consent frameworks require disclosure. Disclosure occurs. Autistic participants remain systematically under-informed. The policy functions as alibi — evidence that something was done, protection against the claim that nothing was offered. The structural failure is invisible because the structure defines what counts as adequate.
The bioethicists recommend individualised consent processes, community advisory boards with autistic membership, disclosure of researcher positionality, and consent materials that speak to the specific concerns of the population being recruited. These are practical steps. But they require researchers to recognise that their default assumptions about what participants need to know may not match what autistic participants actually need to know.
The “reasonable person” was never a neutral standard. It was always a particular kind of person — one whose concerns, values, and frameworks happen to align with institutional defaults. When research concerns a population whose frameworks differ fundamentally from those defaults, consent processes built around the “reasonable person” fail by design.
UK autism research has not addressed this. The question remains open: reasonable to whom?
