Bank-AI collab improves pricing, risk models: Philadelphia Fed's Jagtiani

Bank-fintech collaborations are allowing for more proper pricing and risk underwriting, said Julapa Jagtiani, Senior Economic Advisor and Economist at the Federal Reserve Bank of Philadelphia, on Thursday.

Jagtiani was addressing a webinar on "Beyond Intuition: AI, Crypto, and Social Algorithms Shaping the Future of Finance", organised by Indian Institute of Management Bangalore (IIMB) Centre for Software and IT Management (CSITM).

"When banks team up with vendors of AI solutions, we see more correct pricing of non-prime borrowers. Rather than pricing all as equally risky, AI differentiates who will default and who will not," said Jagtiani.

The webinar was attended by more than 170 participants and included pre-eminent finance professionals from the United States, United Kingdom, and India.

Apart from Jagtiani, the panel comprised Prof. Raghavendra (Raghu) Rau, Sir Evelyn de Rothschild Professor of Finance at Cambridge Judge Business School and an alumnus of IIM Bangalore; and Dr. Marina Niessner, Assistant Professor of Finance at Indiana University's Kelley School of Business.

Niessner noted that financial social media and online narratives significantly contribute to investment decisions more and more.

"We find financial social media to have a powerful influence on trading volumes, often contradicting fundamentals," she said.

But she also warned that access to counterfeit or misleading financial content damages investors.

"It lowers confidence in financial markets as a whole and discourages engagement with genuine content," she added.

Professor Rau gave a critical evaluation of AI models. He warned that modern AI systems, including big language models such as ChatGPT, are really clever statistical machines devoid of actual understanding.

"This inherent trait can cause mistakes or hallucinations, especially dangerous to employ in high-stakes financial judgments," he said.