Coronavirus Reveals Limits of AI Health Tools

The “chatbots” are being marketed as a way to “diagnose” the symptoms. I am not aware of the NHS stance here. I wonder if they effectively triage the symptoms.

Your.md only opens up the list for “practitioners” after asking a few basic questions. The legal ramifications are complex (and it is understandable to label headache as a result of brain tumour). However, it misses on other potential issues,, and each one is trying to peddle their own version of “machine-learning”.

Investors need to be cautioned as this doesn’t represent the advances in medicine. It only represents improved accessibility and an automated system of bringing “patients closer to doctors/health practitioners”.

The U.K. has taken a leading role championing automated health-care tools in recent years. Its state-funded National Health Service, which offers free care for all, has endorsed a number of commercially available chatbots, marketing them to patients in doctor’s offices as a cost-efficient way to triage cases and provide medical advice cheaply for common conditions.

It would be impossible to filter out the “recommendations” for coronavirus in “real-time”. However, the clinical symptoms are broadly known, and the bots have to factor in epidemiological trends to understand the risks of recommending the “diagnosis”. The symptoms resemble common flu which requires specialised testing (and quarantine), which would make things worse in the absence of legislation/workers compensation. What if the sole breadwinner gets it? Will he/she risk being left out without a handout?

via Coronavirus Reveals Limits of AI Health Tools – WSJ