ICD-10-CM/PCS Coordination and Maintenance Committee Fall Meeting: Preview of Coming Attraction Coming Tuesday

Well, thanks, Chuck.

And today, and also since I have a couple extra minutes, I was just going to mention that my highest viewed post of all time on LinkedIn was a post about how to use ChatGPT in order to remove 23,000 emails from my Gmail account to make my Gmail account once more functional.

So chat and cloud can make every employee more capable, but polished output can also conceal the limits, can compose the limits of employees who use AI responsibly for research to analyze, draft, and communicate far faster than colleagues who ignore it.

So in similar roles, competing without AI is becoming increasingly difficult.

A professional supported by AI may summarize a regulation, compare contracts, prepare a presentation, draft computer code, and respond to correspondence in the time another employee needs to complete only one of those tasks.

Healthcare organizations shouldn't choose responsible AI use.

Revenue cycle, reimbursement, compliance, finance, coding, and health information management professionals can use these tools to organize large amounts of information to identify inconsistencies and improve writing and automate repetitive processes and develop better questions for integration.

So experienced professionals can also use AI to explore unfamiliar issues before consulting an appropriate specialist.

Now, the same technology that makes knowledgeable employees more productive can make inexperienced employees appear knowledgeable.

ChatGPT and CLOG produce confident explanations containing technical terminology, calculations, citations, and recommendations.

A staff member with limited understanding of Medicare reimbursement coding clinical documentation statistics or information can therefore create work that looks very authoritative, but appearance is not competence.

An incorrect interpretation can affect claim submission, coding, payment, patient liability, or regulatory compliance.

AI should help qualified employees work faster and examine more possibilities.

It should not authorize unqualified employees to make decisions outside their training credentials or assigned authority.

So verification is the new core competencies.

A general AI policy is not enough.

Managers should identify which tasks may be performed with AI assistance, which conclusions require independent verification, and which decisions must remain under the control of an experienced or credentialed professionals.

Employees should retain primary sources, document important assumptions, and check all calculations and disclose before meaningful AI use.

Organizations need firm rules to protect health information, confidential business information, attorney client communications, and proprietary data.

Information should not be entered into an AI system unless the organization has approved the platform and confirmed that appropriate privacy, security, and contractual protections are in place.

Training should emphasize verification and not merely prompting.

Employees should learn how to test citations and compare AI answers with primary authorities, identify omitted facts and challenge assumptions, and recognize when the question exceeds their own knowledge.

Supervisors should evaluate the reasoning and evidence behind work and not merely polish a final product.

So as far as the competitive professional, AI will not eliminate the need for expertise.

It will make genuine expertise more valuable because experts are always best positioned to recognize when the machine is wrong.