Artificial intelligence simplifies the lives of patients, doctors and hospital administrators. At least, that's the idea. Put in practice, what's the current situation?

‍

Think of an ophthalmologist who needs to diagnose a patient's eye condition.

‍

They can use an AI-assisted machine to gather the data and provide a diagnosis accordingly. But the doctor then comes across a few challenges:

‍

1️. The machine isn't behaving as it's supposed to. Cue a couple hours of troubleshooting, and time is scarce for healthcare professionals.

2️. The doctor also isn't convinced that the AI takes the ethnicity of the patient into account, a legitimate concern for biases.

3️. As a result, the doctor sees clear challenges and sees ways of improving the AI, but there is no feedback loop in place.

4️. The doctor is uncertain of the logic behind the diagnosis and there is no accountability for the machine's decisions.

‍

AI is designed to help our HCPs, but as you can see in the story, there are too many challenges still being faced.

‍

Best practice needs to be implemented from the start so our ophthalmologist can provide the best care and have trust that the AI is working to help provide better care.

‍