SAFE
The core of the job is intraoral hands-on work — bonding brackets, wire adjustments, extraction and expansion decisions, IPR, managing root resorption and TMJ complications — performed on a moving patient over a two-year course of care. AI has already taken real ground on the planning layer: cephalometric tracing, tooth-movement simulation, and aligner staging (ClinCheck-style algorithms) are largely automated, and direct-to-consumer aligner platforms compress the low-complexity case tier. What remains is a licensed clinician who diagnoses skeletal versus dental discrepancies, owns the outcome when a case goes sideways, and physically executes the mechanics.
Tasks largely resist digitisation. Bracket placement on a wet, mobile tooth surface, wire bending chairside, and mid-course corrections when a molar isn't tracking still require a human in the chair, which holds this at 14 rather than 18 — the diagnostic and planning tier that used to be hours of tracing and setup design is now genuinely machine-generated, and Class I crowding cases route to aligner platforms with minimal clinician touch.
Hands-on in uncontrolled environments. Every appointment is intraoral work under retraction — etching enamel, seating archwires, placing TADs, taking impressions or intraoral scans on a gagging patient — and the mouth is an uncontrolled environment with saliva, tongue interference, and a patient who may be an eight-year-old; the 18 rather than 20 reflects that consults, records review, and treatment planning genuinely happen at a screen.
Licensed human required and personally liable. State dental boards license orthodontists individually, board specialty certification gates the title in most marketing statutes, and when root resorption, devitalized incisors, or a failed Class II correction ends in a malpractice claim, the named defendant is the treating orthodontist — no employer or algorithm absorbs that, which is why this sits at the ceiling.
The human relationship is the product. A two-year course of 20-plus appointments with an adolescent and a paying parent makes continuity the product — patients who switch practices mid-treatment describe it as a real loss, and referring general dentists send cases to a specific person, not a clinic — held at 16 rather than 19 because the DTC aligner market proved a meaningful slice of patients will trade the relationship for price.
Exists to be accountable for ambiguous calls. Extract versus expand, surgical versus camouflage in a borderline skeletal Class III, when to stop and accept a compromised result rather than push and risk resorption — these are contested calls with no protocol that settles them, and a 15 rather than 18 reflects that most cases fall into recognized malocclusion patterns with well-documented mechanics.
Has AI actually changed your work?