Ezana and Jay discuss what is machine learning (ML), what are different types of Machine Learning, how Machine Learning is affecting the industries and how we at FCSAmerica can explore this technology to provide better solutions to serve our customers.
Guest: Ezana Beyenne
Host: Jay Pawar
Producer: Tamara Sutton
Editor: Jay Pawar
/
RSS Feed
Good overview, thanks for the information.
Articles like this one, http://time.com/5520558/artificial-intelligence-racial-gender-bias/, show a serious concern for AI’s usage as being potentially unethical when done irresponsibly. Given AI’s vulnerability in this area, do you see AI as a tool in helping us approve loans as we move forward with ELF? If so,what are your thoughts about how we will protect out customers from unfairly being denied a loan?
Tim
Tim – great question! I think that there is always an opportunity for checks and balances when technologies like AI are implemented and as companies learn the power of this new technology. There are also ways to limit just how much learning the machine does based on whether the moderated or unmoderated learning algorithms are used, how much data interpolation is used, etc. I would think we find the sweet spot in our first experiments when we set up an ML system to run parallel with the same data we have available in our existing systems and analyze how the learning system performs and what additional data it might see as significant in making a decision.
To your question about ELF and approving loans, there are regulations that a company must always be able to get back to the factors that were considered when deciding a loan, so I wouldn’t likely see us putting a system in place to decide a loan. Maybe one that considers the data and a decision to be made behind the scenes so we could evaluate if there are other variables we should be considering in our decision engines or scorecards based on the performance of the “behind the scene” model and the models currently used.