Blog post

By 2024, 60% of the data used for the de­vel­op­ment of AI and an­a­lyt­ics projects will be syn­thet­i­cally gen­er­ated

By Andrew White | July 24, 2021 | 0 Comments

synthetic dataSynthetic AssetsDigital TwinData as an AssetData and AnalyticsArtificial Intelligence

A Gartner ‘Predict’, published in the Wall Street Journal this weekend (Fake It to Make It: Companies Beef Up AI Models With Synthetic Data), says: “By 2024, 60% of the data used for the de­vel­op­ment of AI and an­a­lyt­ics projects will be syn­thet­i­cally gen­er­ated”.  The Predict was shared by our very own Er­ick Brethenoux who leads our AI research.

This is a very important Predict because synthetic data has many uses.  The article in this case focuses on fraud and I have already blogged on the use of synthetic data in healthcare; see Re-engineering the Decision – Our Storyline for Data and Analytics. Synthetic data could well touch every industry.

Synthetic data has a bright future if you think about it.  The new set of normals (there may not be a single new normal for some time) organizations will experience going forward, including growth, risk, opportunity and stress, all the same time, triggered a need to re-think how executives and everyone else takes decisions.  This is why our data and analytics storyline is focused on “re-engineering the decision“.  For example, synthetic data can help you cope when decision making brake down when:

  • Estimation or forecast models based on historical data no longer work
  • Assumptions based on past experience fail
  • Algorithms cannot reliably model all possible events due to gaps in real-world data sets

With judicial use synthetic data can help augment new efforts related to using new data sources such small and wide data; see Top Trends in Data and Analytics for 2021: From Big to Small and Wide Data.

But have you even heard of synthetic data?  If you us or are familiar with digital twins you are in a related space.  Digital twins are very structured synthetic (i.e. digital or data) substitutes or “doubles” of physical things used in a digital model.  More broadly synthetic data helps expand and fill out data sets that have gaps or issues.  Only just in January of this year some of our team published a piece of research on the topic: Maverick* Research: Forget About Your Real Data — Synthetic Data Is the Future of AI.  Maverick research is not meant to be a defensible piece of advice built up on years and years of experience and analysis; it is a reach into the future to provoke reaction and new thinking.  Many Maverick ideas die naturally; some come true.  It could well be that synthetic data is coming to a decision of yours very soon!

Comments are closed