Orthoptists

Perform diagnostic tests or measurements, such as motor testing, visual acuity testing, lensometry, retinoscopy, and color vision testing.

AI: Fully automatable - Many ophthalmic diagnostic measurements (visual acuity kiosks, autorefractors, lensometers, automated color testing) are already automated and AI can interpret results.

Provide instructions to patients or family members concerning diagnoses or treatment plans.

AI: Fully automatable - AI can generate clear, evidence‑based, and personalized patient and family instructions for diagnoses and treatment plans at scale.

Prepare diagnostic or treatment reports for other medical practitioners or therapists.

AI: Fully automatable - AI can already synthesize test data and clinical notes into clear, standardized diagnostic and treatment reports suitable for other practitioners and can do so end-to-end when fed the requisite inputs.

Examine patients with problems related to ocular motility, binocular vision, amblyopia, or strabismus.

AI: Partial - AI and eye‑tracking can quantify ocular motility and screen for binocular vision disorders, but comprehensive clinical examination and nuanced interpretation still rely on clinicians.

Evaluate, diagnose, or treat disorders of the visual system with an emphasis on binocular vision or abnormal eye movements.

AI: Partial - AI can provide decision support, pattern analysis, and remote screening for binocular vision and eye-movement disorders but cannot perform the hands-on examinations, nuanced clinical judgement, or therapeutic manipulations required to fully evaluate, diagnose, and treat patients autonomously.

Provide nonsurgical interventions, including corrective lenses, patches, drops, fusion exercises, or stereograms, to treat conditions such as strabismus, heterophoria, and convergence insufficiency.

AI: Partial - AI can prescribe lenses, deliver digital therapy (exercises, stereograms) and guide drop/patch use, but hands‑on fitting and some clinical management remain human tasks.

Interpret clinical or diagnostic test results.

AI: Partial - AI systems can reliably analyze many ophthalmic and orthoptic test outputs (images, eye-movement recordings, measurements) and flag abnormalities, but complex, borderline, or novel cases still require human clinical interpretation and medico-legal oversight.

Develop or use special test and communication techniques to facilitate diagnosis and treatment of children or disabled patients.

AI: Partial - AI can generate adapted test protocols, visual aids, and communication supports for children and disabled patients, yet it cannot fully substitute for the in-person behavioral skills and real-time interpersonal adjustments a clinician provides.

Develop nonsurgical treatment plans for patients with conditions such as strabismus, nystagmus, and other visual disorders.

AI: Partial - AI can propose evidence-based, individualized nonsurgical plans (exercises, prisms, occlusion schedules) and monitor progress, but tailoring, hands-on instruction, and safety oversight by a clinician remain necessary.

Provide training related to clinical methods or orthoptics to students, resident physicians, or other health professionals.

AI: Partial - AI can deliver didactic teaching, simulations, assessment, and just-in-time guidance for trainees, but it cannot fully replace faculty-led hands-on supervision, mentorship, and assessment of clinical skills.

Refer patients to ophthalmic surgeons or other physicians.

AI: Partial - AI can identify indications for referral and generate referral documentation and triage suggestions, but the formal decision, consent, and communication to patients and other clinicians typically require a licensed provider.

Assist ophthalmologists in diagnostic ophthalmic procedures, such as ultrasonography, fundus photography, and tonometry.

AI: Partial - AI can automate image capture guidance, quality control, and interpretation for ultrasonography, fundus photography, and tonometry, but it cannot physically assist procedures or replace the trained operator needed for many diagnostic maneuvers.

Collaborate with ophthalmologists, optometrists, or other specialists in the diagnosis, treatment, or management of conditions such as glaucoma, cataracts, and retinal diseases.

AI: Partial - AI can aggregate findings, generate consult summaries, and support shared decision-making across specialties, but it cannot fully replace human clinical collaboration, negotiation, and responsibility in multidisciplinary care.

Participate in clinical research projects.

AI: Partial - AI can perform literature reviews, design assistance, data analysis, and draft manuscripts for clinical research, but it cannot fully execute patient-facing study procedures, consent, or assume principal-investigator responsibilities without human researchers.

Perform vision screening of children in schools or community health centers.

AI: Partial - AI-driven vision screening devices and apps can detect refractive errors and some pathologies in children, but in-school/community screening still requires human facilitation, behavioural management of young children, and follow-up decisions.

Present or publish scientific papers.

AI: Partial - AI can draft manuscripts, perform literature review, and generate figures or slide decks, but presenting, interpreting novel results, and taking authorship/ethical responsibility cannot be fully automated.