Uploaded July 2026 | Updated November 2026, 4 days ago
AI Hiring Bias: How Transcription Accuracy Decides Your Score
Here is a number from inside the AI hiring machine that should change how you think about video interviews. One major vendor publishes its own word error rate. It sits under ten per cent on average for American English speakers, and it rises to between twelve and twenty two per cent for non-native English speakers, depending on where they are from (Sheard, 2025). Transcription accuracy is not a technical footnote. In algorithmic hiring, it decides whether your words even reach the scoring stage.
This short unpacks why that matters so much for AI recruitment bias. An AI video interview scores a text transcript of your spoken answer. If the words are not captured accurately, the answer cannot achieve a high score, no matter how strong the content was. One recruiter in the research described the effect exactly. At that point the system has stopped assessing communication skills and started assessing English-language ability. Two people can give an equally good answer, and the one the machine transcribes less accurately walks away with the lower score.
Now widen that out, because transcription accuracy does not only vary by accent. It varies with anything the system was not trained to expect. Stuttering. Dysfluency. Atypical prosody. Slowed or uneven pacing. Speech under the kind of processing load that many autistic and ADHD people carry through a high-stakes interview while also managing a countdown timer on screen. AI hiring bias against neurodivergent candidates is often not a dramatic act of exclusion. It is a slow accumulation of small transcription errors that add up to a lower number.
This is also where AI hiring bias and other forms of disadvantage stack. A neurodivergent candidate who is also a non-native English speaker sits at the intersection of two transcription penalties at once. The research is clear that these systems are largely built and trained offshore, then sold worldwide, so the norm being applied to job seekers in Aotearoa New Zealand was often shaped somewhere else entirely.
In this short I explain what transcription accuracy actually is, why it is doing so much hidden work in algorithmic hiring, and what you can reasonably ask an employer about the format before you record. A few examples worth knowing. You can ask whether the interview is scored by an algorithm or reviewed by a person. You can ask for a written alternative to a recorded video. You can ask for extra time, or for the timer to be removed where speed is not a genuine requirement of the job. None of these requests are unusual, and each one is easier to make once you understand that transcription accuracy, not just the quality of your answer, is shaping your score. The free toolkit gives you the legal grounding across five jurisdictions and the self-advocacy scripts to ask with confidence.
AI Hiring Bias: How Transcription Accuracy Decides Your Score
Here is a number from inside the AI hiring machine that should change how you think about video interviews. One major vendor publishes its own word error rate. It sits under ten per cent on average for American English speakers, and it rises to between twelve and twenty two per cent for non-native English speakers, depending on where they are from (Sheard, 2025). Transcription accuracy is not a technical footnote. In algorithmic hiring, it decides whether your words even reach the scoring stage.
This short unpacks why that matters so much for AI recruitment bias. An AI video interview scores a text transcript of your spoken answer. If the words are not captured accurately, the answer cannot achieve a high score, no matter how strong the content was. One recruiter in the research described the effect exactly. At that point the system has stopped assessing communication skills and started assessing English-language ability. Two people can give an equally good answer, and the one the machine transcribes less accurately walks away with the lower score.
Now widen that out, because transcription accuracy does not only vary by accent. It varies with anything the system was not trained to expect. Stuttering. Dysfluency. Atypical prosody. Slowed or uneven pacing. Speech under the kind of processing load that many autistic and ADHD people carry through a high-stakes interview while also managing a countdown timer on screen. AI hiring bias against neurodivergent candidates is often not a dramatic act of exclusion. It is a slow accumulation of small transcription errors that add up to a lower number.
This is also where AI hiring bias and other forms of disadvantage stack. A neurodivergent candidate who is also a non-native English speaker sits at the intersection of two transcription penalties at once. The research is clear that these systems are largely built and trained offshore, then sold worldwide, so the norm being applied to job seekers in Aotearoa New Zealand was often shaped somewhere else entirely.
In this short I explain what transcription accuracy actually is, why it is doing so much hidden work in algorithmic hiring, and what you can reasonably ask an employer about the format before you record. A few examples worth knowing. You can ask whether the interview is scored by an algorithm or reviewed by a person. You can ask for a written alternative to a recorded video. You can ask for extra time, or for the timer to be removed where speed is not a genuine requirement of the job. None of these requests are unusual, and each one is easier to make once you understand that transcription accuracy, not just the quality of your answer, is shaping your score. The free toolkit gives you the legal grounding across five jurisdictions and the self-advocacy scripts to ask with confidence.






![Autism Sleep Solutions: New Hope for Adults
Autism Sleep Solutions: New Hope for Adults
For years, autistic adults with insomnia were offered almost nothing backed by research. That is finally starting to change.
A 2026 review in the journal Sleep described autistic adult sleep as a case of high prevalence, low evidence, urgent need. Between 2011 and 2021, no sleep intervention for autistic adults could be rated as evidence-based, and early pilot studies had just 10 and 8 participants (Richdale & Morris, 2026). Now the first randomised trial of an online insomnia programme adapted for autistic adults, called i-Sleep, has reported that 38% of participants improved, compared with 11% of controls, and 35% no longer had insomnia at the end. More people dropped out than usual, which tells us the format needs to fit autistic people as well as the content.
In this Short, I explain what cognitive behavioural therapy for insomnia is, why an adapted version matters, and what to ask a clinician: have they worked with autistic clients, and can the programme be adjusted for literal instructions, sensory needs and pace?
For leaders and managers: poor sleep in autistic staff is often linked to sensory load and anxiety at work and at home. Predictable schedules, quieter workspaces, fewer surprise changes and a clear fatigue risk conversation can make a real difference to wellbeing and safety at work.
I would love to hear from you in the comments. What gets in the way of sleep for you, and what has actually helped? Your experience helps other people feel less alone, and it shapes the videos I make next.
This Short comes from Part 2: Autism and Sleep: Why Autistic Adults Never Feel Rested, one of three videos in my neurodivergent sleep series. Watch the full video here: [FULL VIDEO LINK]. Grab the free printable brain dump and two-week sleep log here: [DOWNLOAD LINK]. The whole series, covering ADHD, autism and AuDHD sleep, is in this playlist: [PLAYLIST LINK]. Subscribe for more research you can actually use.
I am Nicola Knobel, a late-diagnosed AuDHD woman and a health, safety and risk professional based in Aotearoa New Zealand. I make research-based videos for neurodivergent adults and the leaders who support them, and I always show you where the evidence comes from, and where it is still thin.
This video is general information and is not medical advice. If sleep is affecting your health, work or safety, please talk to your GP. In New Zealand, melatonin is a restricted medicine, so any conversation about it belongs with your doctor.
#Shorts #AuDHD #NicolaKnobel #CBTI Autism Sleep Solutions: New Hope for Adults](https://i.ytimg.com/vi/j6FlP8npxTo/mqdefault.jpg)



