Machine Learning and Technology: the future of health care?

The role of Artificial Intelligence (AI) and Machine Learning in healthcare were hot topics at a panel discussion that took place at The Wearable Tech Expo, held 8 March 2017 at the Excel Centre in London.

The panel focused on what role machines can realistically play in real life situations. Diagnostics seemed to be the answer that was most readily provided – if good data is fed into machines than probability scores and validation processes should allow them to assist doctors in accurately diagnosing patients. At the moment the human error rate for diagnostics is 7-9% and the consensus is that this could be reduced with the aid of intelligent machines that have been thoroughly tested and validated.

The reality is that machines are already playing a successful role in diagnostics. Currently skin cancer is more effectively diagnosed by machines and machines used in radiology are so far advanced that at some point radiologists may see their roles reduced. AI is playing a role in addressing mental health issues – for example, social media sites are being developed that help predict and pick up subtle changes in behaviour patterns. A similar technology is also being created to help diagnose dementia.  These types of programmes can make a diagnosis based on the changes and patterns in the types of language and tones used in texts.

The increased presence of AI and technology in medicine will certainly raise questions. The panel spoke about legal issues such as those of liability in the case of misdiagnosis that will have to be addressed before it becomes the norm. Blame and causation will also need to be defined and ethical issues will begin to appear as use of the technology becomes more widespread.

Despite some issues and confusion that still exist around AI and Machine Learning, the general feeling at the Expo was that they will start to play a more significant role in diagnostic medicine and in healthcare more generally over the next ten years.