After reviewing federal regulators’ traditional theory of redlining, we discuss the types of underwriting practices that are likely targeted by Director Chopra’s recent comments expressing concern about “algorithmic redlining,” examine how the use of machine learning (ML) underwriting models incorporating alternative data can be more inclusive than traditional logistic regression models and result in more approvals for protected class members and “credit invisibles,” and offer our thoughts on actions that technology and credit providers should take in response to Director Chopra’s comments when developing and using ML models.… Continue Reading
machine learning
This week’s podcast: A deep dive into the federal agencies’ request for information on artificial intelligence, with special guest Nicholas Schmidt, Chief Executive Officer, SolasAI
After discussing the current state of the regulators’ knowledge about artificial intelligence and machine learning (ML) in underwriting models, we examine the regulators’ key areas of focus for ML models (explainability/accuracy in adverse action notices, potential hidden bias, testing for disparate impact), discuss how to test for and counteract disparate impact and how to search for less discriminatory alternatives in ML model development, and consider regulators’ possible next steps.… Continue Reading
This week’s podcast: Using machine learning and artificial intelligence in debt collection
In this podcast, we are joined by Scott Ferris, CEO of Attunely, a provider of machine learning (ML) and artificial intelligence (AI) technology to the debt collection industry. We look at how changes in consumer behavior have impacted collections, technology’s role in collections, state law’s/GDPR’s impact on ML/AI and compliance strategies, how ML/AI can improve profitability, and perceived impediments to adopting ML/AI.… Continue Reading
CFPB gives boost to use of alternative data and machine learning
A new CFPB blog post titled “An update on credit access and the Bureau’s first No-Action Letter” provides a boost to lenders using alternative data and machine learning in their underwriting models.
The Bureau issued its first (and so far only) no-action letter in September 2017 to Upstart Network Inc.… Continue Reading