Orthoptists deal with how the two eyes work together. They measure ocular deviation, test binocular vision and fusion, and set therapy plans that a child or an adult follows for months. People asking whether AI will replace orthoptists are usually thinking of automated vision screening devices and image-reading software. Those tools are real, and they are changing parts of the work. The honest story here is task erosion, not a job disappearing.
Why the exam room still needs an orthoptist
Most orthoptic measurement is a negotiation with a patient. A cover test with prisms needs steady fixation, a cooperative gaze and a clinician who can tell a true deviation from a tired or distracted eye. Software can log a number. It cannot coax a four-year-old into looking at a target, notice that the child is guessing, and repeat the test a different way.
Therapy is the second anchor. Planning and supervising orthoptic exercises, patching regimens and fusion training runs over weeks. Adherence is the hard part, and adherence is a human relationship with a parent as much as with a patient. An app can remind. It cannot adjust the plan when a family says patching is causing fights at school.
The job also sits inside a clinical team. Orthoptists prepare measurements that surgeons rely on for strabismus planning, and they explain findings to families in words those families can use. Both tasks carry responsibility that someone has to hold. That is why our coverage score, Can AI do it?, sits where it does: 17 out of 100.
What software handles, what it assists, and what stays with people
The share of task time our model puts fully on AI is 5%. It covers the paperwork end of the role: drafting visit notes from structured measurements, and pulling prior results into a progress summary before a review appointment. Nothing diagnostic sits there on its own.
Assisted work is larger, at 23% of task time. Automated photoscreeners and digital alignment devices capture raw numbers faster than hand testing, and image-grading tools flag retinal findings for a clinician to confirm. The orthoptist still decides whether the reading is trustworthy and what it means for this patient.
Work that needs a person is 72%. That group holds the hands-on tests, the therapy sessions, the counseling of parents about amblyopia treatment, and the judgment calls when test results and the child in front of you disagree. The task list above shows which tasks fall where.
What the evidence actually covers
Our evidence grade for this occupation is D. That means no study has tested an AI system against qualified orthoptists on orthoptic assessment, so we publish no parity number. Published work in eye care has focused on image-based screening, mainly diabetic retinopathy and other retinal disease, rather than on measuring and treating binocular vision disorders.
A grade that low is a gap in the research, not a verdict. Two kinds of study would settle it. First, a prospective comparison on the same patients: automated alignment and amblyopia detection against orthoptist measurement, blinded, with error margins reported for young children. Second, a trial of software-led therapy monitoring against clinician-led review, measured on visual outcomes and dropout. Until something like that exists, our quality parity measure, Is it better than a person?, stays ungraded by number.
The market picture is steadier. BLS puts employment in this group at 28,630 and median pay at $115,210, with projected growth of 5.5% from 2025 to 2035 (BLS, 2025). That is a small, specialized workforce growing slowly, not one under visible pressure.
When the picture could change
Most likely after 2042 (8 in 10 of our scenarios). Our replacement-year method explains what that window is measuring.
Two things could pull it earlier. Automated vision screening is already cheap to run next to staffing a clinic, and the cost panel on this page shows how wide that gap is. If pediatric screening devices get good enough to triage referrals reliably, fewer routine assessments reach an orthoptist, and entry-level posts are usually the first to thin out.
Two things hold it back. Over half of the work has a physical element, and the robotics tier for that part is a dexterous humanoid, which is hardware that does not exist in clinics today. Credentialing and clinical accountability are the other brake: someone qualified signs off on a measurement that leads to surgery, and that rule changes slowly.
How to stay needed as an orthoptist
Lean into the tasks that sit in the human group. Pediatric assessment with uncooperative or pre-verbal patients. Therapy design and adjustment over a long course. Pre-surgical measurement and the conversation with the surgical team about what the numbers can and cannot support.
Two skills raise your value either way. One is device literacy: knowing how a photoscreener or grading tool reaches its output, and where it fails, so you can overrule it with reasons. The other is teaching, with families, trainees and the technicians who run the equipment.
What to do: get fluent with the screening devices in your own clinic before anyone asks you to supervise their results.
Nearby roles move in the same direction. Compare this page with Optometrists, Ophthalmologists, Except Pediatric and Ophthalmic Medical Technologists, or put any two side by side in the job comparison tool. For the wider picture, see the healthcare diagnosing and treating practitioners family, the healthcare sector page and our list of jobs that mostly need a person. Our scoring is published in full at how we score jobs.