Many of the lines of artificial intelligence research today bear on the health sciences. An example concerns the right to privacy. Is it possible to develop a comprehensive database about the health conditions of the human species without infringing on the privacy rights of the individuals who compose it? What if Jane Smith doesn’t want the world to know when she contracted Covid and how severe her case was? Does respecting her rights come at the cost of comprehensive data?
One line of AI research involves “synthetic data.” The term refers to the creation of an artificial database that mimics the real world, approximating original data and retaining its key statistical correlations. For many purposes the synthetic data works as well as the original — in predictive value, for example, and in suggesting fruitful lines of continued research. Yet once a proper synthetic database is created the original can be erased and no threat to privacy rights need arise.
In Pill Form:
Roche, one of the pharmaceutical giants, is using synthetic medical data to guide its clinical research, and the German insurance company Provinzial is using such data for its predictive analytics. The field is still considered to be in its infancy.