Washington:
The U.S. financial industry faces a “huge” data gap between large and small banks as it deploys artificial intelligence to fight fraud, the Treasury Department said on Wednesday, noting that smaller institutions are at a disadvantage.
The U.S. Treasury Department said that while larger banks have more internal data to develop artificial intelligence models to prevent fraud, the same is not true for smaller banks.
The U.S. Treasury Department said there was a need to close the gap, noting “insufficient data sharing among businesses.”
The latest recommendations and reports come after President Joe Biden unveiled an executive order on regulating artificial intelligence in October, and the Treasury Department is now taking steps to identify risks and challenges.
The order directs federal agencies to develop new security standards for artificial intelligence systems while requiring developers to share their security testing results and other critical information with the U.S. government.
“Artificial intelligence is redefining cybersecurity and fraud in financial services,” said Nellie Liang, U.S. Treasury Undersecretary for Domestic Finance.
Liang added that the Treasury report lays out a vision for how financial institutions can “securely plan their lines of business and disrupt rapidly evolving AI-driven fraud.”
Cybersecurity information sharing is mature, but “little progress has been made to enhance fraud-related data sharing,” the report said.
It said the U.S. government could help build a “data lake of fraud data” that could be used to train artificial intelligence.
The department also called for “labeling” that would allow the department to clearly identify what data is used to train models for AI systems provided by vendors and where they come from.
Other steps identified by the Treasury Department include “explainability solutions” for advanced machine learning models and greater consistency in defining artificial intelligence.
(Except for the headline, this story has not been edited by NDTV staff and is published from a syndicated feed.)