Resolved to learn about AI and what it can do for your business in 2019? These 10 resources will take less than an hour of your time
What is big data and AI and how can I explore this field?
You’ve heard of “artificial intelligence” and “big data.” You know it’s a big deal, but you might not know how to leverage it to your business advantage. You might not know what questions to ask or what “right” looks like when talking with vendors, consultants or even internal teams.
I’m here to help you get a handle on artificial intelligence (AI), machine learning (ML) and advanced analytics – all things big data — in 2019 and beyond. Here are the top 10 articles, blogs and podcasts I recommend to clients. Each requires five minutes or less of your time.
1. Oracle Big Data Blog
Both the Oracle Big Data Blog and R-bloggers do a great job of explaining what artificial intelligence is. Is it the same as machine learning, data science and advanced analytics? You’ll learn that AI is any method that allows a computer to emulate human behavior. Machine learning, which is a way of teaching a computer without programming an explicit set of rules, is one of those methods.
3. Harvard Business Review: “Artificial Intelligence for the Real World”
What should you tackle first with AI? This article references a recent study of more than 150 AI projects and “reveals that highly ambitious moon shots are less likely to be successful than ‘low-hanging fruit’ projects that enhance business processes.”
This reflects my experience, too. “Low hanging fruit” projects have distinct advantages: narrower in scope and cost as well as more achievable. These attainable projects help build momentum and support within your management team for future endeavors.
4. NPR: “Will Using Artificial Intelligence To Make Loans Trade One Kind Of Bias For Another?”
5. NPR: “Could Your Social Media Footprint Step On Your Credit History?”
If you’re a lender or you sell products based on the credit of your customer or assess risk of consumers or small businesses, you should learn more about the role of alternative data.
After all, character is usually the first thing a lender wants to understand about a borrower. When asked if commercial credit is based primarily on money or property, John Pierpont “J.P.” Morgan once replied, “No, sir, the first thing is character.”
Lenders historically use credit scores as a proxy for character. Big data — or more accurately alternative data — is increasingly being used as an additional lens to understand someone’s character. There are incredible risks to this. But there are also great opportunities, because other data, such as income and FICO scores are biased.
These two NPR “All Tech Considered” podcasts discuss machine data and alternative data for lending in a fair and balanced way. Check out “Will Using Artificial Intelligence To Make Loans Trade One Kind Of Bias For Another?” and “Could Your Social Media Footprint Step On Your Credit History?”
6. The Washington Post: “Mortgage algorithms found to have racial bias”
7. 2River blog post: “AI in 2019 for Lending and Financial Services Industry”
One of the biggest risks is that AI and ML reflect the same biases that people have had for decades. AI is supposed to remove human bias. As reported in The Washington Post — “Mortgage algorithms found to have racial bias” — and elsewhere, if not done properly, lending algorithms can have significant racial bias.
While there are technical approaches and modeling building techniques that companies should use to avoid creating an algorithm with racial bias, companies should also ensure they have diversity in their AI teams.
For more on how this cultivate different viewpoints and experiences, check out our related blog post. More on this to come as 2River will be speaking on this topic in March at American Banker’s Retail Banking 2019 conference in Austin, Texas.
8. Harvard Business Review: “Getting Value from Machine Learning Isn’t About Fancier Algorithms — It’s About Making It Easier to Use”
One of the challenges for companies that start an AI or ML project is that while the technology works, but they struggle to use it. For example, if ML can tell a fitness club which members will leave in 90 days, what should the gym do with that information? If ML tells a lender which borrowers are most likely to default in the next quarter, what should that lender do?
This is where your data science team needs to be closely tied to your business and operations teams. My team and I have found that front-line managers have techniques to reengage customers. We have found that loan officers and relationship managers have ways to determine if a borrower’s underlying issue is either a willingness to pay or a capacity to pay — and then use “carrot” and “stick” strategies to manage the recovery of funds.
This article shares stories of how organizations get value from ML, and the authors have deep experience working with AI technology.
9. The Economist: “Babbage: When an algorithm decides your fate”
AI and ML are often rightly criticized for being a black box and rightly so. This podcast covers the challenges and steps organizations take in explaining ML. Fast forward to the 11th minute.
10. IBM Blog: “John Foreman: Data science is equal parts technology and translation”
Ready to hire your first data scientist or your first chief data scientist? Read this IBM Q&A with my former colleague, John Foreman, describing his experiences joining MailChimp as its first chief data officer. Particularly insightful and relevant is his comment that, “The main hurdle to doing data science in an established company is really change-management. It’s not technical.”
Bottom line: You need to be an informed decision maker as you determine if and how AI, ML and advanced analytics are right for your business this year and in the years to come.
If you want to start exploring predictive analytics, sign up for a FREE LIFT account now.
Not ready, learn more in our resource section.