AI Resilience Report forHealth Information Technologists and Medical Registrars

Many health records tasks are already done by computers or AI helpers. For instance, patient data coding is often aided by computer-assisted coding (CAC) tools. These programs use artificial intelligence (like natural language processing) to read doctors’ notes and suggest standardized codes ^{[1] [1]}.

Hospitals report this can speed up work (one study saw ~20% faster coding ^[1]), but human coders still review the results. Experts note that fully-automated coding is still “promising” but not ready to replace people ^{[1] [1]}. Other tasks rely on computer systems too: almost all hospitals use Electronic Health Records (EHRs) now ^[1], so retrieving a patient chart often means a quick database search.

Software can flag missing information or rule violations, but staff do the final check for accuracy. Security is also handled with both tech and humans: encryption and AI can spot unusual access, but staff enforce strict privacy laws (like HIPAA) to keep records safe ^[1]. Finally, assigning patients to DRG billing groups is usually done by billing software, though coders still confirm those groupings.

In summary, many tasks are supported by AI or software, but the most critical work — checking details, ensuring privacy, and making judgments — still needs human oversight ^{[1] [1]}.

Hospitals are increasingly trying AI in health information jobs, but change is gradual. A 2025 survey found 80% of U.S. health systems are exploring or piloting AI tools for billing and coding ^[2]. They hope AI will save time and money in the long run – even one report estimates about $150 billion in annual healthcare savings from key AI applications ^[1].

Demand for health information workers remains strong (job openings are growing about 7% over the decade) ^[3]. However, hospitals face barriers. The biggest issues are cost and integration: new AI systems can be expensive to buy, install, and train staff on ^{[2] [2]}.

Smaller hospitals, which have tighter budgets and fewer IT staff, tend to adopt AI more slowly ^[2]. Privacy and safety concerns also slow AI use: strict patient-data rules (HIPAA) and complex regulations make it harder to share and process health records for AI development ^{[4] [1]}. In practice, many places start with AI as a helper (for example, suggesting codes) while people make final decisions.

Experts stress that human skills remain vital – workers with medical knowledge, attention to detail, and ethics are needed to guide AI and use data responsibly ^{[1] [1]}. In short, AI is becoming part of this field, but it’s mostly augmenting workers, not replacing them, and people will still be in charge of the most important parts of the job.