The Future of Identity in Financial Services

My testimony, today, represents my own views as a public interest technologist. As a computer scientist and organizational scholar with expertise in open government data, I am part of a growing movement of people using STEM skills in non-profits and the public sector. My academic specialty is understanding the organizational dynamics that shape the production and consumption of information, especially in organizations that have a public mission. Artificial intelligence is not infallible. Even the most successful artificial intelligence systems used by online financial platforms require human input. For Americans to participate equally in our financial system, we need inclusive innovation that is aware of difference. Ignoring AI exceptions in financial services risks excluding many in our society because they are outliers from expectations. Organizations must begin to think about how they will handle future disputes over AI errors. Artificial intelligence in the financial sector is an ethical, mathematical, and policy issue. To illustrate this, I will elaborate on three main points: 1. Artificial intelligence produces errors. When operating “at scale” even low error rates can impact millions. Errors in financial services will be consequential to specific individuals. 2. Because organizations are more likely to believe their technology systems over the experiences of individuals, individuals need procedures for recourse in the event of processing error. 3. Systems built to consider a broader range of populations must be more fault tolerant of cultural difference to be robust. Artificial intelligence, often implemented to save labor costs, will still require human labor to handle anticipated exceptions. A dispute resolution process solves two problems: procedural justice and technology improvement. First, it establishes a procedure to preserve the sanctity of human experience in situations where organizations may be more likely to trust the AI over a customer. Second, it provides the necessary feedback for incremental improvement of the technology. Artificial intelligence will have its exceptions. People need procedures to assert the authority of their lived experience over the authority of the numbers.