Bill Harvey: Baye’s watch

Use of algorithms and machine learning should be embraced. Abuse must be avoided. We are likely to hear a lot more in the coming years about the application of artificial modelling in clinical practice. With a very limited resource budget, anything that improves efficiency of clinical decision making should be welcomed. It should not be assumed, however, that artificial intelligence is a threat to our livelihoods.

Algorithms used in online shopping predict our patterns of behaviour to target advertising and marketing and they do this effectively. Issues of communication and ethics are less well served. You may remember the story of the teenager in the US whose father discovered she was pregnant when questioning the amount of maternity related spam she was being sent as a response to her online activity. Decisions about access to loans or insurance cover are sometimes based on analysis of online profiling and have been known to throw up patterns of decision-making both inappropriate and selectively biased towards a white male affluent programmer environment.