As vice president of enterprise machine learning at TD Bank, Matt Fowler faces the difficult task of overseeing the bank’s efforts to ensure its algorithms are free of bias. Fowler will talk about his work on a panel discussing bank-fintech collaborations at Bank Innovation Build, which runs from November 6-7.
According to Fowler, keeping bias out of machine learning algorithms is an exercise in due diligence. “When you remove data that could present bias in the model’s output — for example, men, women or address — the models are so smart that they can introduce biases through other information [considered], including where you shop, what you buy, etcetera,” he said.
Fowler has worked at TD Bank for more than 10 years. Before that, he was a senior director of business process control at CIBC. An important lever to fight bias in AI and machine learning is to have people from different backgrounds working on the algorithms.
“Diversity in the team that builds AI models is key,” Fowler said. TD acquired an AI company, Layer 6, in 2018. According to Fowler, 90% of that team is from outside Canada. “Diverse backgrounds, diverse approaches and [diverse] thought processes gives us a better end product.”
Bank Innovation Build will take place in Atlanta on November 6-7. Other agenda items include implementing next-generation customer experience; speeding up development times at banks; designing the bank of the future; user experience design; and selling innovation internally. Register here.