
Can disabled people…? What autocomplete reveals about views on disability
Inspired by Safiya Umoja Noble’s Algorithms of Oppression, this post explores a neglected area of autocomplete research. While stereotyping in search suggestions has been studied for race, gender, and other attributes, disability is largely absent from the literature. I report the results of a small audit of disability-related autocomplete queries across major search platforms.
After a recent reread of Safiya Umoja Noble’s excellent Algorithms of Oppression, I went down a rabbit hole looking for empirical studies that audit search engine autocomplete for stereotyping. Noble shows how search engines are not neutral mirrors of public interest, but systems that reproduce and amplify social hierarchies through design, ranking, and monetisation choices. As she puts it:
Search results reflect the values and norms of the search company’s commercial partners and advertisers and often reflect our lowest and most demeaning beliefs, because these ideas circulate so freely and so often that they are normalized and extremely profitable.
A surprising gap emerged: although multiple researchers have documented stereotyping and dehumanisation in autocomplete suggestions related to race, gender, sexuality, religion, and age, disability appears to be largely absent from this literature. That absence is notable, given how often disabled people are subjected to questions that challenge their basic capacities, legitimacy, and even humanity.
It was too enticing a research gap to resist. In this post, I describe the results of a small, manual audit of autocomplete suggestions across four major search platforms using disability-related queries.
How does autocomplete question personhood?
The first query template I used was “can [disabled people]…” These questions tend to frame disabled people’s lives as matters of capacity and permission. They ask whether disabled people are able to think, feel, work, have relationships, reproduce, or participate in civic life at all. In practice, these queries frequently function as tests of basic personhood.
To make sense of the patterns in these suggestions, I grouped them into six domains of personhood that the questions implicitly challenge:
Inner life: whether disabled people think, understand, or experience emotions in recognisably human ways (e.g. intelligence, understanding, empathy, falling in love).
Subjective experience: whether disabled people can perceive, feel, or sense the world as others do (e.g. sensation, pain, sensory perception).
Moral character: whether disabled people are dangerous, deceptive, or morally suspect (e.g. violence, aggression, manipulation).
Reproductive fitness: whether disabled people are permitted or able to have sex, reproduce, or become parents (e.g. sex, pregnancy, children, parenthood).
Adult life: whether disabled people can live independently, work, or sustain ordinary adult roles (e.g. employment, independent living, relationships).
Civic participation: whether disabled people can participate fully in civic and legal life (e.g. voting, jury service, military service, legal responsibility).
The figure below summarises how often different domains of personhood are questioned in autocomplete suggestions. For each disability term, it shows the proportion of autocomplete suggestions that fell into each domain.
For clarity, the figure shows results from Google and Bing only. These two platforms produced the most complete and consistent sets of autocomplete suggestions across disability terms. Yahoo and DuckDuckGo are shown in the table below.
Several patterns stand out:
- Different disabilities attract different kinds of personhood doubt. Cognitive and developmental conditions are most likely to prompt questioning of inner life and adult independence, while psychiatric conditions attract suspicion about moral character.
- Physical and sensory disabilities are more often framed through perception and sensation, driven by curiosity about pain, dreams, and bodily experience rather than legitimacy or moral worth.
- Questions about reproduction cut across many disability types, suggesting that assumptions about who should have children remain deeply embedded in public thinking about disability.
These patterns are remarkably consistent between Google and Bing.
For completeness, the table below shows the suggestions for each disability on each of the four search engines, with colours used to indicate terms belonging to each of the six domains.
“Can [disabled people]…” autocomplete suggestions by disability and search engine
| 'Can [disabled people]...' suggestions by disability and search engine | ||||
| Inner Life Subjective Experience Moral Character Reproductive Fitness Adult Life Civic Participation | ||||
| Disability | Bing | Yahoo | DuckDuckGo | |
|---|---|---|---|---|
| disabled people | get married, vote, drive, join the military, park on double yellows, work, park on double yellow lines, have kids, get food stamps, own guns | have sex, join the military, vote, work, drive, get medicaid, park on double yellows, get married, get medicare, own guns, have kids, get out of jury duty | go to jail | join the military, vote, work, drive, get medicaid, park on double yellows, get married |
| blind people | dream, see in their dreams, drive, cry, see in dreams, have schizophrenia, get schizophrenia, see light, see, use echolocation | dream, drive, cry, see, see light, be schizophrenic, blink, swim, write, play video games, hallucinate, use echolocation | dream, see their dreams, what see | dream, drive, cry, see, see light, be schizophrenic, blink, swim |
| deaf people | drive, speak, hear their thoughts, read, drive in australia, drive in victoria, hear themselves, hear, get tinnitus, hear their own voice | speak, drive, hear their thoughts, read, hear, drive legally, talk with their voice, drive uk, listen to music, get a drivers license, hear with hearing aids, have tinnitus | drive, read, talk, hear their thoughts, go to the normal olympics, drive cars, read english | speak, drive, hear their thoughts, read, hear, drive legally, talk with their voice, drive uk |
| paralysed people | get pregnant, walk again, drive, have kids, feel pain, talk, swim, feel, poop, feel their legs | feel sex, feel pain, have kids, feel touch, get hard, get pregnant, feel their legs, ever walk again, feel their genitals, feel, talk, blink | — | feel pain, have kids, feel touch, get hard, get pregnant, feel their legs, ever walk again |
| wheelchair users | fly, swim, drive cars, walk, use escalators, use uber, drive, go on roller coasters, ice skate, go on cruises | drive, have sex, reach counter height | use, get, see, drive | drive, reach counter height |
| people with cerebral palsy | drive, talk, walk, have children, understand you, live a normal life, have kids, work, talk normally, swim | walk, drive, speak, have kids, have sex, swim, have children, drink alcohol, work, live independently, talk, get married | get, take, live, run | walk, drive, speak, have kids, swim, have children, drink alcohol |
| people with dyslexia | read, learn to read, drive, write, join the military, be smart, spell, have high iq, read music, be successful | drive, learn to read, go to college, claim pip, read numbers, join the military, read, read minds, be smart, speak, have kids, write | read | drive, learn to read, go to college, claim pip, read numbers, join the military |
| people with intellectual disability | drive, go to college, work, vote, live on their own, be smart, get married, drive a car | drive, vote, live independently, work, have children, get married, have kids, get disability | get, work, take, buy | drive, vote |
| people with learning disabilities | go to college, vote, join the military, be smart, live on their own, go to college for free, go to university, have high iq, drive, have children | drive, work, live independently, have children, vote, have kids, be successful, get married | get, use, play, drive | drive, work |
| people with Down syndrome | how long live, drive, have children, have kids, vote, live independently, walk backwards, reproduce, drink alcohol, be smart | have kids, drive, vote, walk backwards, drink alcohol, reproduce, consent, be smart, go to college, live alone, have autism, have babies | drive, have children | have kids, drive, vote, walk backwards, drink alcohol, reproduce, consent, be smart |
| autistic people | drive, be empathetic, have kids, say the r word, be smart, be extroverted, hear electricity, join the army, fall in love, be social | join the military, drive, have kids, lie, have sex, be narcissistic, work, be violent, join the army, be smart, hear electricity, live on their own | drive, fall in love, can play professional soccer, can participate in the olympics, what be less sensitive to | join the military, drive, have kids, lie, be narcissistic, work, be violent |
| people with ADHD | join the military, drive, meditate, multitask, read books, be organised, have manic episodes, have sensory issues, be successful, stim | drive, join the military, multitask, get pip, be successful, have sex, work, hyperfocus, be smart, focus, be hypnotized, meditate | join the military | drive, join the military, multitask, get pip, be successful, work, hyperfocus |
| people with schizophrenia | drive, work, live a normal life, get ndis, smoke weed, have kids, live alone, live normal lives, be cured, own guns | work, drive, live normally, be violent, get ssi, be cured, own guns, be dangerous, join the army, get disability, drive uk, function | get, see, take, feel | work, drive, live normally, be violent, get ssi, be cured, own guns, be dangerous |
| people with bipolar | work, live without medication, hear voices, take antidepressants, drive, be dangerous to others, live a normal life, be violent, drink alcohol, have kids | live normal lives, work, drive, hallucinate, work, be dangerous, take antidepressants, get disability, get guns, take adderall, take trazodone, be dangerous | — | live normal lives, work, drive, hallucinate, work, be dangerous, take antidepressants, get disability |
| people with fibromyalgia | work, donate blood, exercise, drive, have kids, get tattoos, donate plasma, get a blue badge, get massages, get pip | work, get disability, take statins, donate blood, have kids, have children, donate plasma, drink alcohol, take o | get, feel, go, give | work, get disability, take statins |
| people with chronic fatigue | work, drive, exercise, drive, donate blood | exercise, get disability, donate blood, drink alcohol, work, donate plasma, live long, drive | cause, symptoms, qualify | — |
What stereotypes do search engines surface?
Having examined what capacities people question, I next wanted to explore to what traits they assume using a second query template: “why are [disabled people] so…”. This query is designed to elicit different kinds of assumptions and stereotypes. These questions tend to presuppose a stereotyped trait or behaviour, so they more often surface stereotypes about character, danger, emotional volatility, or otherness.
I grouped the resulting suggestions into six categories:
- Negative stereotype: moral judgments, (e.g. angry, mean, lazy, manipulative, annoying, difficult)
- Pity: framing disabled people as fragile or suffering (e.g. tired, weak, vulnerable, anxious, depressed, scared, sensitive)
- Savant: assumption of compensatory special talents (e.g. good at math, good at piano)
- Strong: strong or physically strong
- Happy: infantilising positivity framing (e.g. happy, cheerful, joyful, positive, content)
- Positive stereotype: other benevolent stereotyping (e.g. kind, nice, creative, smart, funny, honest, sweet)
Unlike the earlier queries, these frequently returned no suggestions at all, particularly on Google and DuckDuckGo. This could be due to filtering or potentially volume of queries, though one would imagine Google would have at least as much to work with as Bing. For this reason, the table below presents the results directly rather than summarising them in a figure.
| 'Why are [disabled people] so...' suggestions by disability and search engine | ||||
| Negative Stereotype Pity Savant Strong Happy Positive Stereotype | ||||
| Disability | Bing | Yahoo | DuckDuckGo | |
|---|---|---|---|---|
| disabled people | strong, physically strong, happy, funny, nice, smart | — | — | — |
| blind people | good at singing, smart, good at piano, nice | — | stupid, mean, angry, tired | — |
| deaf people | strong, expressive, nice, cute | — | bad, stupid, hard, angry | — |
| paralysed people | — | tired, emotional, weak, numb, strong, happy, vulnerable, scared | mean, angry, stupid, happy | — |
| wheelchair users | — | vulnerable to injury, vulnerable to falls, tired, tall, vulnerable to injuries, vulnerable to accidents, emotional, important | expensive, bad, angry, high | — |
| people with cerebral palsy | happy | emotional, sensitive to light, smart, sensitive, sensitive to noise, intelligent, tired, special | tired, angry, mean, old | — |
| people with dyslexia | — | smart, good at math, good at writing, sensitive to light, hard to read, hard to understand, good at reading, emotional | angry, tired, bad, low | — |
| people with intellectual disability | — | — | hard, much | — |
| people with learning disabilities | — | sensitive to light, emotional, successful, smart, special, sensitive, intelligent | angry, tired, bad, hard | — |
| people with Down syndrome | happy, strong, kind, nice, flexible, physically strong, sweet, affectionate, loving, confident | happy, nice, strong | happy, much | happy, nice, strong |
| autistic people | smart, honest, flexible, physically strong, nice, funny, creative, happy, good at one thing, smart reddit | strong, physically strong | smart | strong, physically strong |
| people with ADHD | funny, creative, smart reddit | tired, sensitive, impulsive, good at math, emotional, hyperactive, sensitive to noise, sociable, lazy, smart | bad, hard, annoying, common | tired, sensitive, impulsive |
| people with schizophrenia | — | — | difficult, much | — |
| people with bipolar | — | emotional, moody, depressed, angry, sociable, mean, sensitive, tired | much | — |
| people with fibromyalgia | tired | tired, sensitive to pain, emotional, bloated, skinny, cold, sensitive, angry | tired, bad, angry, low | — |
| people with chronic fatigue | — | tired, irritable, sensitive to pain, emotional, anxious, weak, sensitive, cold | bad | — |
Several observations are notable:
- Google returns lots of positive stereotypes like happy, strong, and creative.
- In contrast, Bing returns mostly negative stereotypes (people with bipolar get a particular bad run: emotional, moody, depressed, angry, mean).
- Yahoo also returns a lot of negative stereotypes, with terms like stupid, mean, bad, annoying, and difficult all coming up for most disabilities.
- Down syndrome consistently attracts happy across all platforms and nice across three.
- Chronic illness terms only attract pity (tired, weak, anxious).
- Strong and physically strong appear with surprising frequency across disabilities. For example, autistic people get ‘physically strong’ for three of the four search engines.
The contrast between Google and Bing here is particularly striking given how similar their results were for the earlier “Can…” queries. This suggests different filtering or moderation approaches to stereotype-based prompts.
Does terminology change the questions asked?
If autocomplete reflects underlying assumptions about personhood, the language used to describe disability might be expected to shape those assumptions.
There’s ongoing debate about whether to use identity-first language (“autistic people”, “disabled people”) or person-first language (“people with autism”, “people with disabilities”). Autistic and Deaf communities generally prefer identity-first language, while person-first remains standard in many clinical and policy contexts. And of course, individuals have their own preferences.
This led me to wonder if the choice of language changes what suggestions are offered. To explore this, I compared autocomplete suggestions for identity-first and person-first language across four conditions: disability (generic), deafness, autism, and bipolar disorder. Only Google results are shown, as other platforms frequently returned no suggestions.

Three findings stand out:
- For disability and deafness, identity-first and person-first language produce broadly similar patterns.
- “Autistic people” attracts substantially more questioning of inner life than “people with autism” but less questioning of capacity for adult life.
- “People with bipolar” attracts more moral character suspicion than “bipolar people”.
These results complicate the assumption that person-first language consistently humanises how disability is perceived.
The table below shows the suggestions for each disability and framing for Google, with colours used to indicate terms belonging to each of the six domains.
“Can [disabled people]…” Google autocomplete suggestions by disability and framing
| Google 'Can [disabled people]...' suggestions by disability and framing | ||
| Inner Life Subjective Experience Moral Character Reproductive Fitness Adult Life Civic Participation | ||
| Disability | Identity-first | Person-first |
|---|---|---|
| disability | get married, vote, drive, join the military, park on double yellows, work, park on double yellow lines, have kids, get food stamps, own guns | join the military, go to jail, vote, drive, get married, have kids, work, live on their own, own a gun, go to heaven |
| deaf | drive, speak, hear their thoughts, read, drive in australia, drive in victoria, hear themselves, hear, get tinnitus, hear their own voice | drive, speak, read, from birth speak, talk, and blind communicate, and mute drive, hear, have tinnitus, from birth talk |
| autism | drive, be empathetic, have kids, say the r word, be smart, be extroverted, hear electricity, join the army, fall in love, be social | drive, join the military, live a normal life, lie, work, join the army, drink alcohol, have kids, live independently, get married |
| bipolar | live a normal life, be narcissistic, drink alcohol, donate blood, know they are bipolar, tell if they’re bipolar, join the military, control their emotions, work, tell they are bipolar | work, live without medication, hear voices, take antidepressants, drive, be dangerous to others, live a normal life, be violent, drink alcohol, have kids |
Final thoughts
Autocomplete suggestions reveal something about how the public thinks about disabled people (or at least, what questions they type into search engines). The patterns are not random. Psychiatric conditions attract suspicion about dangerousness. Cognitive disabilities prompt doubts about inner life and capacity for adult independence. And across nearly all conditions, people ask whether disabled people can or should have children at all.
What stood out most to me was how often stereotypes surfaced as ostensibly positive. Terms like strong, physically strong, happy, and nice appeared frequently. This was particularly noticeable for autistic people on Google: smart, honest, physically strong, nice, funny, creative, happy. As an autistic person, this does not reflect my experience of how autistic people are typically perceived in everyday life, which is far more often shaped by assumptions of incapacity, social difficulty, or fragility.
These findings are interesting in their own right, but it is also notable how little effort this analysis required. Manually collecting the data and analysing it took under a day. That such a basic study appears to be missing from the literature reflects how research priorities in the area of algorithmic bias have largely excluded disability.
Search engines reinforce the questions people already ask about marginalised groups. For disabled people, those questions often start from a presumption of incapacity, danger, or otherness. Until disability is treated as a core dimension of algorithmic fairness research, these patterns will continue to go unexamined.