There’s been a lot of talk about AI (Artificial Intelligence) recently. This is more so after ChatGPT debuted in the public domain, quickly followed by other similar applications.
Some have been bothered by AI taking over and replacing humans in some job categories. Think of recent strikes in the US by those working in the entertainment industry. But the potential gains emanating from AI cannot be understated.
The integration of AI in the medical industry isn’t new. There are AI applications in diagnostics that are contributing to better patient outcomes. A good example is in radiology (medical imaging), where AI applications have eased the human workload, and sometimes surpassed human interpretations.
In fact, almost all facets of healthcare can benefit from some form of AI integration and contribute to benefits all around. But the downsides of AI in medicine are also apparent. A good example has been magnified by a lawsuit in the US about the use of AI in medical insurance approval processes.
In the said lawsuit, the insurance company has been accused of using AI algorithms to deny lots of medically necessary claims.
The AI system apparently processes claims within 1.2 seconds and rejects them without any human ever physically reviewing patients' records. The claimants think human reviewers would not necessarily have rejected some of the claims.
What hasn’t been brought into prominence in the lawsuit is how the AI algorithm has eased the processing of claims and increased efficiency. There must be gains in turnaround times for claims, lower needs for manpower, and possibly overall savings for the insurance company. Further, down the line, physicians involved in case reviews may have been redirected to other more humanistic tasks.
- When the spine fails to develop fully: exploring spina bifida
- Cliniva: A solution helping women get care that puts their experience first
- Early warning systems can help build resilience against future flooding
- We want Kenyans to reach doctors from where they are
Presumably, any savings may eventually accrue to lower insurance premiums for patients in the long run. What cannot be in contention is that humans continue to play a major role in medicine. Algorithms driving AI medical applications must initially be coded by humans before machine learning takes over.
And humans must continue to monitor AI-driven outcomes and be ready to override AI decisions that may be entirely incorrect. Eventually, conflicts between AI and humans may fade into oblivion. And we may never know when that moment actually comes!
Dr Alfred Murage is a Consultant Gynecologist and Fertility Specialist. [email protected]