Anthem has partnered with Google Cloud to create a synthetic data platform that can better detect signs of fraud.
Synthetic data copies real-world observations and are used to train algorithms in place of organic information, which may lack volume, quality and variety. The insurance provider is seeking to develop large amounts of this data to teach its algorithms to identify fraudulent claims, as well as abnormalities in patient health records. It then plans to deploy them to assess real-world data from its members to better protect their privacy and in the long-term, help personalize their care, reported The Wall Street Journal
It has enlisted Google Cloud to use deep generative models to produce approximately 1.5 to two petabytes of synthetic data, which will be made up of artificially generated medical histories, healthcare claims and other key items. “More and more, synthetic data are going to overtake and be the way people do AI in the future. The variation of the data is going to be very, very important,” said Anthem’s chief information officer Anil Bhatt in a statement.
Synthetic data can reduce biases in real-world data sets and provide companies with faster access to the information they need to test and validate algorithm efficacy. It does, however, at times create data sets that are worse than real ones.
Anthem already uses AI algorithms to detect fraud and abuse in insurance claims and expects the new synthetic data to scale these solutions and personalize care and identify when patients may require medical intervention.
The company partnered with Google Cloud last year to enhance its data analytics and AI capabilities to become more customer-centric and help members throughout their entire healthcare experience. “Synthetic data models, in our opinion, will ultimately fuel the promise of what big data can deliver,” said Chris Sakalosky, managing director of U.S. healthcare and life science at Google Cloud.