Future of AI
No doubt AI is coming and will impact the future of healthcare in a significant manner. What I believe will eventually differentiate the AI providers of the future is not so much their respective deep learning, “machine learning” algorithms but how they leverage and aggregate complementary technologies and information to bring added value to medical practice.
There are numerous AI companies all working to develop learning algorithms for specific abnormalities in medical imaging. Over a period of time, the initial sensitivity and specificity of these algorithms are expected to improve as the more subtle and the more indeterminate abnormalities are identified and fed back into the learning network for retraining yielding improved performance results.
Since access to the full scope of imaging findings and abnormalities will become increasingly available to all developers, it is reasonable to assume that everyone’s algorithms will converge to very similar performance levels. That being the case, it will be important to recognize how other technologies and information can complement AI to help improve effectiveness and efficiency. AIE represents such a technology. By improving the clarity and detail of findings flagged as abnormal or suspicious, AIE will serve as a powerful enhancement to AI.