Medical billing and coding is often sold as a stable, work-from-home career, and it is also often named as a job AI will automate. Both claims have some truth. The data shows a job with high task exposure but growing demand, and a role that is shifting from entering codes toward checking and fixing them.
The short answer: AI is likely to change medical coding a lot, but current evidence does not point to the job disappearing. The ILO's 2025 index places medical records and health information technicians in its significant-exposure band. The US Bureau of Labor Statistics projects employment of medical records specialists to grow 8 percent from 2025 to 2035, while noting that AI-powered coding tools may affect demand. No projection is certain.
How exposed coding and billing work is
How to read the table: the mean task score runs from 0 to 1, and the band is the ILO's grouping, from not exposed through minimal exposure to gradient 4, the highest (ILO Working Paper 140, 2025). More on the method in will AI take my job?
| Occupation (ISCO-08) | ILO category | Mean task score (0 to 1) |
|---|---|---|
| Medical Records and Health Information Technicians | Gradient 3 | 0.52 |
| Medical Secretaries | Gradient 3 | 0.53 |
| Statistical, Finance and Insurance Clerks | Gradient 4 (highest) | 0.64 |
| Accounting and Bookkeeping Clerks | Gradient 4 (highest) | 0.64 |
| Data Entry Clerks | Gradient 4 (highest) | 0.70 |
Gradient 3 means a significant share of tasks is exposed, but not all. The pure data-handling roles nearby, such as data entry and insurance clerks, sit in the highest band. The more a billing role looks like data entry, the more exposed it is; the more it involves reading clinical records and resolving problems, the closer it sits to coding.
What projections expect
The BLS projects employment of medical records specialists, the group that assigns diagnosis and procedure codes for patient care, statistics and billing, to grow 8 percent from 2025 to 2035, much faster than average, with about 14,000 openings a year. It links the growth to an aging population and more chronic conditions such as heart disease and diabetes. It also says that the increase in adoption of AI-powered solutions that make medical coding more efficient may affect demand for these workers (BLS Occupational Outlook Handbook).
So the volume of records to code is rising, while each coder can handle more of it. Which effect wins will vary by employer.
Which tasks AI is taking on
- Suggesting diagnosis and procedure codes from clinical notes.
- Checking claims for missing fields and common errors before submission.
- Sorting and routing denied claims.
- Drafting appeal letters and patient billing messages.
- Extracting data from scanned documents and forms.
What still needs a person
- Reviewing suggested codes: NIST lists confabulation, confidently stated false output, among the main risks of generative AI (NIST AI 600-1). A wrong code can mean a denied claim or an incorrect bill.
- Complex and unclear records: unclear documentation, multiple conditions and unusual procedures.
- Querying clinicians when the record does not support a code.
- Denials and appeals that need judgment and follow-up with payers.
- Compliance and audits: making sure coding follows the rules and that someone is accountable for it.
- Patient data care: health records are sensitive, and AI tools must be used within the organization's privacy rules. See AI agents and personal data.
How the job is changing
For many coders, the work is moving from typing codes to auditing them: the AI suggests, the coder confirms, corrects or escalates. That is a typical human-in-the-loop role. The BLS notes that employers may prefer certified medical records specialists, and people who combine coding knowledge with comfort checking AI output are best placed. Billing roles that are mostly data entry face the most pressure.
If you are considering the field because of "work from home" ads, be careful: see our guide to remote AI jobs for common job scam signs.
Read next: will AI take my job?, checking AI output and which jobs are turning digital collar.
Pay and how to get in
The US Bureau of Labor Statistics puts the median annual pay for medical records and health information technicians at $51,140 in May 2025, against $50,980 for all workers. The typical education needed to enter is: postsecondary nondegree award. See how this job compares in AI exposure by pay and AI exposure by education level.
Questions and answers
Will AI replace medical coders?
Not on current evidence. The BLS projects 8 percent growth for US medical records specialists from 2025 to 2035, though it notes AI coding tools may affect demand. The work is shifting toward reviewing AI-suggested codes.
How exposed is medical coding to AI?
The ILO places medical records and health information technicians in gradient 3, significant exposure, with a mean task score of 0.52.
Is medical billing more at risk than coding?
Billing work that is mostly data entry is more exposed: the ILO puts data entry and insurance clerks in its highest exposure band. Billing that involves denials, appeals and payer follow-up needs more judgment.
What will medical coders do as AI spreads?
Review and correct AI-suggested codes, handle complex records, query clinicians, manage denials and appeals, and support compliance and audits.
Is medical billing and coding still worth going into?
The BLS projects about 14,000 openings a year for medical records specialists and faster than average growth. Employers may prefer certified candidates, and skills in checking AI output are increasingly useful.