A recent article described a patient’s frustrating experience with a medical misdiagnosis and argued that artificial intelligence may eventually perform some diagnostic tasks more effectively than humans.
As often happens in discussions about AI, the conversation quickly gravitates toward replacement. Will AI replace doctors? Therapists? Teachers? Lawyers?
I suspect that is often the wrong question.
Personally, I would not want an AI serving as my therapist. I would not want one performing surgery without human oversight. In both cases, human judgment matters too much.
What I would welcome is an AI that helps those professionals perform their jobs more effectively.
Imagine walking into a doctor’s office after an AI-assisted intake process. Before the appointment even begins, the system has already gathered symptoms, identified inconsistencies, flagged risk factors, and suggested follow-up questions for the physician. During the examination, it continues acting as a second set of eyes, surfacing possibilities the practitioner might want to consider.
The physician remains responsible for diagnosis and treatment decisions.
The AI simply helps ensure that important signals are not overlooked.
That strikes me as a far more realistic and valuable future than fully autonomous medical decision-making.
The same principle applies in education.
I don’t want AI replacing faculty members.
I do want AI helping instructors identify struggling students earlier, improve feedback processes, generate alternative explanations for difficult concepts, and reduce administrative burdens that pull them away from teaching.
In many fields, the most effective future is probably not human versus machine.
It’s human plus machine.
AI excels at processing information, finding patterns, and maintaining consistency across large volumes of data.
Humans excel at context, empathy, ethical judgment, and understanding nuances that fall outside predefined parameters.
The most interesting innovations will likely come from combining those strengths rather than forcing a choice between them.
As I’ve said elsewhere, good technology rarely succeeds by replacing expertise.
It succeeds by making expertise more effective.
