Welcome to the Research Assistant Weekly Newsletter - a subscriber-only resource for insight into emerging compliance challenges, details on peer calls, and links to new Research Assistant reports, documents, tools, and more.
|
Sponsored by TCN
|
Generative Artificial Intelligence is the new buzzword across the industry, and this week Sara Woggerman spoke on a panel to discuss the risks associated with this and how to proactively prepare your Compliance Management System. Here are some of the highlights from our discussion.
Let’s start with understanding the difference between machine learning, artificial intelligence, and generative artificial intelligence.
-
Machine Learning is a pathway to artificial intelligence. ML uses techniques to train algorithms to automatically learn insights and recognize patterns from data, applying that learning to make increasingly better decisions.
-
Artificial Intelligence refers to systems designed to respond to a particular set of inputs. These systems have the capability to learn from data and make recommendations, decisions or predictions based on that data.
-
Generative AI is a class of AI models that can create or emulate something new. Generative AI models are trained on a set of data and learn the underlying patterns to generate new data that mirrors the training set. This can include images, videos, audio, text, and other digital content.
Generative AI is quickly advancing but what are some of the things we need to consider before we use these tools within our organization.
First, you may want to consider restricting open AI sources from employees. Nearly all of the open source options available disclose that everything you add to it becomes their information. The unintended consequence of using a tool like this is that confidential or sensitive information may be shared as a result. ARM companies should be drafting policies, procedures, and controls to manage who has access to these tools, and determine when it’s acceptable to use. You may also want to require employees to disclose when they use it for work related projects, reports, or content.
You should consider the viewpoint of regulators and how they are looking at generative AI tools. The CFPB has been putting us on notice for the last 18 months about the use of generative AI in financial services. Here are the cliffnotes version of what we can expect them to do in the near future.
-
The use of AI in credit decision making – We should expect to see new disclosure requirements when AI is used in making decisions based on credit worthiness, or in the ARM space why we are treating an account a specific way.
-
Decision tree evidence – This will undoubtedly be added to the CFPB’s examination procedures. Which means we will need to be able to evidence how the data was used, and the way it was used to determine an outcome.
-
Risk assessments on algorithmic bias- Lookback and risk assessments will be required to ensure that we are not unintentionally running afoul of our legal obligations.
Generative AI is a fascinating tool that can help us become more efficient in our jobs but just like all new technologies it’s critical to ensure we understand the technology we’re using, implement robust policies, and procedures, and effectively evaluate the risks of consumer harm.
Documents and Crowdsourced Materials:
Top Reads:
Upcoming Webinars/ Other Announcements:
|