SAFE
The measurement half of the job — refraction, visual field testing, retinal photo screening for diabetic retinopathy and glaucoma suspects — is already being handled by autorefractors, OCT, and FDA-cleared AI image readers, and remote/kiosk refraction keeps chipping at the routine exam. What holds is the licensed, hands-on part: slit lamp and dilated fundus exam, contact lens fitting and follow-up, diagnosing and treating ocular disease with prescription drugs, and signing the prescription and referral decision. Modal optometrist is in retail or private practice doing mostly routine vision exams, so volume pressure is real even though the role itself isn't going away.
Dipped in 2020, then grew past where it started.
Median pay $115,250 → $136,570 -5.2% in real terms
This line is counted by the Bureau of Labor Statistics — the one figure on this page that isn't a judgement of ours. Headcount moves on demand, offshoring, demographics and the business cycle, and automation is one term among several, often not the loudest.
So a falling line is not evidence that AI did it, and a rising one is not evidence that it won't. Both happen in this register: some occupations resist automation and shrink anyway, others are highly automatable and keep growing. The marked year is 2020.
BLS projection, 2024–2034
+8% 47,800 → 51,600 on the projections basis
Hard to automate, and growing
The work resists current AI and the BLS projects +8% more of these jobs by 2034. Note that safe does not mean well paid — several of the fastest-growing resistant occupations are among the lowest paid on the register.
Different clocks. The score is what current AI could do to this work today. The projection is how many of these jobs will exist in 2034. Everything between the two — how fast employers actually adopt, whether demand grows in the meantime — is why they can point opposite ways without either being wrong.
~2,400 openings a year on average, including replacing people who leave.
OptometristOptometry Doctor (OD)Therapeutic Optometrist
Holding it up: liability shield . Weakest point: task resistance .
Mixed — a routine tier and a judgment tier Autorefractors and phoropter automation plus IDx-DR-class AI already cover the refraction and diabetic-retinopathy screening that fills much of a retail chair, but slit-lamp biomicroscopy, gonioscopy, dilated fundus exam, punctal plug insertion, and fitting scleral or multifocal lenses across follow-up visits still need a clinician's hands and eyes at the lamp — hence mid-band rather than the 6 a pure refraction technician would earn.
Hands-on in uncontrolled environments You are physically at the slit lamp, applying tonometry probes and diagnostic drops to a moving human eye, inserting and removing lenses on-eye, and removing corneal foreign bodies — real patient contact in an exam lane, but a lane you control with fixed equipment and a chin rest, which is why this sits at 14 and not with the field paramedic at 18.
Licensed human required and personally liable Optometry is state-licensed with a defined scope of practice, and you personally sign the spectacle/contact prescription and the therapeutic drug order and carry the malpractice exposure for a missed melanoma, papilledema, or angle-closure — no supervising physician absorbs that call, which is what puts this near the ceiling.
Exists to be accountable for ambiguous calls The consequential calls are yours alone and often ambiguous: is that disc a physiologic cup or early glaucoma, does this red eye get an antibiotic or a same-day retina referral, is the amblyopia still treatable — you make them on one visit's data without imaging consult, which is discretion with vision loss on the line.
The verdict above describes this occupation as a whole. Almost nobody does the typical version of a job — tick what's actually in your week and see how your own mix sits.
Your task mix speaks to task resistance (12/20 here) — how much of the day's work current AI already does. That is the dimension the boxes above are about.
It cannot move the other three. Liability shield (18/20) is whether the law requires a licensed human to sign. Trust premium (14/20) is whether buyers specifically pay for a person. Judgment and accountability (14/20) is whether the role exists to own consequential calls. Those are facts about the occupation's standing, not about which tasks are in your week — a paralegal who does only trial exhibits still holds no licence. Together they are 46 of this occupation's 72 points (64%).
Embodiment (14/20) is also a property of the work rather than the worker, but we don't tag individual tasks as physical or not, so the picker can't tell you anything about it. That's a limit of this tool, not a claim.
Did we get the list right? Tell us what's missing — the tasks are written from the outside, and you're reading this from the inside.
The moves above are yours to make. This is the other half: what would have to change in the world for the occupation itself to score higher. None of it is in any one person's gift, but it is where the floor actually comes from. Scores here are not a one-way ratchet. Only two of the five dimensions — task resistance and embodiment — track what machines can do. The other three track law, what buyers will pay for, and who is answerable, and those move in both directions, often in response to the same pressure AI creates. If every lever below landed, this occupation would score around 82/100, still SAFE.
Continued expansion of optometric scope-of-practice laws to include laser procedures (YAG, SLT) and minor lid surgery — enacted in Oklahoma, Kentucky, Louisiana, Alaska, Mississippi, Wyoming, Virginia, Colorado — moves the role from refraction toward owning treat-or-refer calls and surgical outcomes with personal liability.
State optometry boards tightening the rule that an autonomous AI/telerefraction device cannot generate a spectacle or contact lens prescription without a licensed optometrist's in-person exam within a defined window — the pattern already enacted in states restricting online eye exam apps (e.g. South Carolina's Eye Care Consumer Protection Act, Indiana's telerefraction restrictions) and pushed by the AOA. Also FDA/board rules requiring a licensed reader to countersign autonomous diabetic retinopathy screening (IDx-DR/LumineticsCore) referrals before treatment.
Task-mix shift is genuine here: if kiosk/remote refraction and autonomous retinal AI absorb the routine screening tier, the residual day is dilated fundus and slit lamp exam, complex contact lens and keratoconus/scleral fitting, dry eye and glaucoma management, and pediatric/binocular vision therapy — all judgment- and hands-heavy. This raises the fraction of remaining work AI cannot do while shrinking total demand per patient.
Growth of the medical-eyecare share of practice (in-office punctal plugs, foreign body removal, IPL/meibomian gland expression, laser procedures where scope permits) increases the share of the day requiring physical manipulation of an uncooperative, moving eye.
The limit. Trust premium has no realistic upward route: retail chain and vision-plan channels commoditize the exam, and patients choose on insurance network and appointment availability, not on the individual optometrist. Also note the score can hold while headcount demand per practice falls — resistance of the role is not the same as volume of jobs.
| New York-Newark-Jersey City, NY-NJ | 2,580 | $161,030 +18% |
| Los Angeles-Long Beach-Anaheim, CA | 2,360 | $138,940 +2% |
| Houston-Pasadena-The Woodlands, TX | 1,250 | $124,640 -9% |
| San Francisco-Oakland-Fremont, CA | 1,180 | $142,580 +4% |
| Chicago-Naperville-Elgin, IL-IN | 1,160 | $145,630 +7% |
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 890 | $153,980 +13% |
| Dallas-Fort Worth-Arlington, TX | 870 | $140,790 +3% |
| Phoenix-Mesa-Chandler, AZ | 860 | $115,960 -15% |
| La Crosse-Onalaska, WI-MN | 40 | $175,810 +29% |
| Asheville, NC | 40 | $167,980 +23% |
| Seattle-Tacoma-Bellevue, WA | 420 | $167,810 +23% |
We have no reported case of a named organisation automating this occupation. Not one deployment, not one announcement.
That is worth saying out loud next to a score of 72. The verdict above is about what the work exposes — what current AI could do to these tasks. It is not a claim that anyone has done it. For this occupation those two things have come apart completely: the capability argument is on this page, and the evidence column is empty.
Has AI actually changed your work? One tap, anonymous, and the running tally is public. Nothing else is asked of you.
Rather than check back: get the digest and we'll tell you what changed — or watch a single occupation from its own page.