Will AI replace health information technologists?
Part of the work, not the whole job. AI could take on about 59% of the working time in a typical health information technologist job today. The other 41% still needs a person.
41% of the work still needs a person.
The bracket under the bar is the evidence range for the share AI could take on: 44 to 59%. One end comes from how well AI did each responsibility in real use, or a second rating where that isn't measured. The other is our figure, which also follows research across the industry.
15 responsibilities in a typical health information technologist job. Median pay $68,020 a year in the U.S. (BLS, May 2025).
Still needs a person
- Clarify unclear codes with doctors
- Train medical records staff
- Promote privacy and security awareness
- Protect medical record confidentiality AI helps
- Design secure healthcare databases AI helps
- Code and abstract patient data AI helps
AI can do or help with
- Maintain procedures and system queries AI helps
- Code and abstract patient data AI helps
- Run health record indexing systems AI helps
- Protect medical record confidentiality AI helps
- Develop staff training materials AI helps
- Recommend healthcare system upgrades AI helps
The biggest responsibilities on each side, ranked by a person's time on the left. "AI helps" means AI takes on part of the work, or most of it with a person checking what it produces, so one can show on both sides. See all 15, with the reason for each and the cost of handing them to AI.
Is this your job?
Upload your résumé to see your Human Edge: how much of this job your experience holds in the 41% that still needs a person. You also get a free plan with what to learn next. We never keep your résumé.
Questions people ask
Will AI replace health information technologists?
Part of the work, not the whole job. AI could take on about 59% of the working time in a typical health information technologist job today. The other 41% still needs a person. That is a split of the work, not a forecast of jobs lost.
Which parts of the job can AI do?
The biggest ones AI can do or help with: maintain procedures and system queries, code and abstract patient data, run health record indexing systems, protect medical record confidentiality, develop staff training materials, recommend healthcare system upgrades.
Which parts still need a person?
The biggest ones: clarify unclear codes with doctors, train medical records staff, promote privacy and security awareness, protect medical record confidentiality, design secure healthcare databases, code and abstract patient data.
How is this measured?
The U.S. Department of Labor lists 16 responsibilities for this job (O*NET 31.0), and the split covers the 15 people in it do today. Each is scored on 4 questions: can AI do it, does it get it right, must a person sign for it, and do people need to deal with a real person. The split is weighted by how much of the job each responsibility takes up. The evidence range for the share AI could take on is 44 to 59%. One end comes from how well AI did each responsibility in real use, or a second rating where that isn't measured. The other is our figure, which also follows research across the industry. Scores version 0.7.17, updated October 28, 2026.
Responsibilities from O*NET 31.0 (U.S. Department of Labor). Pay from the Bureau of Labor Statistics, May 2025. Scores are calibrated to published research and updated every month. How it works.
Related jobs
- Will AI replace judicial law clerks? 42% needs a person
- Will AI replace paralegals? 48% needs a person
- Will AI replace patient representatives? 54% needs a person
- Will AI replace pharmacy technicians? 59% needs a person
- Will AI replace radiologic technologists? 63% needs a person
- Will AI replace family doctors? 66% needs a person