■ MODERATE RISK ■ Healthcare
No — but the job is being rebuilt around smarter machines. Ventilators are learning to manage themselves; coaching a panicking patient through a BiPAP mask remains defiantly human.
“AI-controlled ventilators adapt in real-time. But setting up the scared patient is on you.”
Our AI replacement risk score — how we score jobs
Respiratory therapists live where medicine meets machinery. A shift might include titrating ventilator settings in the ICU, running weaning trials, drawing and interpreting arterial blood gases, giving nebulizer treatments on the floors, fitting CPAP masks on patients who hate them, responding to codes as the airway person, and teaching a newly diagnosed COPD patient how not to end up back in the hospital next month.
The machines are getting opinionated. Closed-loop ventilation modes already adjust support breath-by-breath, and AI systems increasingly recommend settings, predict which patients will fail extubation, and flag deterioration before the numbers look bad. Blood gas analysis is automated; interpretation is heading the same way. The routine-treatment circuit — the scheduled nebulizer rounds that fill much of a floor shift — is exactly the kind of protocolized work hospitals love to compress, delegate, or automate away.
What survives is everything that happens at the bedside when physiology and fear collide. A ventilator can optimize its own pressures, but it cannot notice the patient biting the tube, reposition an airway, talk a claustrophobic person into keeping the mask on, or make the judgment call during a code when the monitor and the patient disagree. RTs are also the ones who catch that the 'ventilator problem' is actually a kinked circuit or a pneumothorax. With respiratory disease rising and hospitals chronically short-staffed, our risk score of 45 reflects a role where the cognitive monitoring shrinks, the caseload per therapist grows, and the hands-on and crisis work keeps the profession alive — busier, and leashed to smarter equipment.
Automatability: our editorial assessment of current and near-term AI capability
Closed-loop ventilation and AI-assisted weaning are entering ICUs now, and by the early 2030s much of the settings-adjustment work will be machine-led with human oversight. Expect staffing ratios to stretch — more patients per RT — rather than mass layoffs. Through 2040 the profession transforms toward crisis response, complex cases, and patient-facing work, buoyed by an aging population that keeps generating respiratory disease.
Yes, with eyes open. Demand is strong — aging populations produce respiratory disease faster than schools produce RTs — and the hands-on core of the job resists automation. But expect the work to shift: fewer hours adjusting ventilator knobs, more patients per therapist, more emphasis on critical care, emergencies, and education. Specialize early and the outlook is solid.
They'll eliminate some of the knob-turning, not the therapist. Closed-loop modes handle routine titration well, but ventilated patients generate non-routine problems constantly — secretions, agitation, circuit failures, sudden decompensation — that require hands and judgment at the bedside within seconds. The realistic outcome is each RT overseeing more ventilators, not hospitals running them unattended.
Critical-care and emergency airway competence, comfort managing AI-assisted equipment (including knowing when it's wrong), and genuine patient-communication ability. The therapists most exposed are those doing only routine floor treatments; the most protected combine ICU-level clinical judgment with the soft skills that keep terrified patients compliant with therapy.
Routine ventilator titration, blood gas interpretation, and predictive monitoring — the data-driven tasks — are automating now. Scheduled nebulizer rounds may be compressed or delegated. Last to go: code response, non-invasive ventilation coaching, difficult airways, and education. If your shift is mostly the first list, start moving toward the second.