Witnesses before the House Financial Services Committee Task Force (the "Task Force") on Artificial Intelligence provided testimony on mitigating risks associated with the deployment of artificial intelligence ("AI") technologies. The witnesses cautioned against overreliance on AI solutions and emphasized the need to preserve digital identities in an increasingly connected digital world.

NYU Steinhardt School Assistant Professor of Data Policy Anne Washington warned that AI technologies remain incapable of disambiguating large data sets, and stressed that even the most advanced AI systems will continue to require human input for the foreseeable future. Professor Washington suggested balancing AI technologies, such as customer identification and fraud detection, with a human-based dispute resolution process to (i) establish a procedure that preserves the value of human experience under circumstances where organizations are more likely to defer to the AI over the customer, and (ii) continuously gather feedback for incremental improvement of the technology.

Accenture Security Managing Director Valerie Abend made AI-based cybersecurity recommendations, including (i) ensuring federal or state legislation addressing AI is "technology-neutral" and does not choose winners and losers, (ii) creating policies to protect and advance AI innovation, thus harnessing this technology to help financial institutions defend against cyber threats, and (iii) adopting a national data privacy law to help foster an "effective [and] robust" digital identity ecosystem to exist alongside AI technology.

Better Identity Coalition Coordinator Jeremy Grant recommended improving the identity ecosystem within the private sector by:

  1. focusing on the development of the next-generation remote identity proofing and verification systems;

  2. reconsidering how Social Security Numbers are used in the United States;

  3. encouraging and prioritizing the use of strong authentication;

  4. seeking ways for the United States to coordinate with other countries and harmonize requirements; and

  5. being aware of new identity solutions and how best to utilize them through consumer and business education.

Turnkey Risk Solutions President and Founder Amy Walraven highlighted the "widespread and inconceivably pervasive" threat of synthetic identity fraud on a global scale. She stated that in order for institutions to detect and remediate synthetic identity fraud, it must be (i) universally defined and (ii) incorporated into the United States control framework.

SecureKey Technologies Chief Identity Officer Andre Boysen emphasized the importance of digital services that provide privacy and control over personal information for individuals, while simultaneously offering identity validation claims from several parties in a single transaction. Mr. Boysen stated that positive results can be achieved only through (i) acceptance and trust of the service from the individuals involved and (ii) the ability to reach a large user base quickly.

Commentary

Joseph V. Moreno

While AI is being touted by FinCEN and others as an exciting and potentially game-changing technology to help fight cyber fraud and identity theft, it remains a long way off from replacing human efforts altogether and presents its own risks to financial institutions. That is the big takeaway from the Task Force hearing.

AI solutions permit the automation of existing routine manual tasks. But AI software's ability and security is only as good as the programmer who designed it. An over dependence on new technology with few industry standards can be costly. The methods behind how AI technology makes decisions are often unclear, potentially leaving financial institutions unable to assess and validate their decision-making or justify them to regulators. There are also reports that AI algorithms can introduce bias and discrimination into decision-making, prompting significant ethics and policy concerns. AI appears to be the way of the future for the financial sector. How it will be implemented and what principles and controls will govern it, however, will determine whether embracing AI will be a success or a failure.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.