Socure has released a new report that suggests that automated identity technology can help mitigate bias and raise financial inclusion. In the report, Socure compared the performance of its own identity verification platform to that of a legacy customer identification program (CIP).
The results showed that Socure could automatically process a much higher number of potential users. In that regard, the company produced a 99 percent increase in the auto-approval rate for those under the age of 25, as well as 46, 36, and 28 percent increases for Asian, Hispanic, and Black populations, respectively. The company also reported low levels of bias, claiming that its solution performed well for all demographics and that there were no statistically significant differences for any of the top 200 Sigma Identity predictors evaluated in the study. That remained true across demographics like race, gender, age, class, and immigration status.
As it stands, there are more than 45 million credit-invisible people in the United States, defined as those who do not have a credit file or have a file that does not have enough information for a credit score. That includes immigrant populations and young populations that are just entering the workforce, as well as those from marginalized backgrounds.
Socure argued that its technology could double the number of people from those groups that could be auto-approved for financial services in the US on an annual basis. That, in turn, would make it much easier for marginalized groups to participate in the financial system. The tech can also help mitigate the unfair biases that emerge when one set of rules is applied to everyone. Such policies often have a disproportionate impact on disadvantaged populations, who are forced to jump through more hoops simply because they do not fit the default profile.
“No Machine Learning models at Socure get deployed without rigorous performance, sensitivity and fairness bias testing,” said Socure Chief Product and Analytics Officer Pablo Abreu. “Socure continuously monitors and evaluates production Sigma Identity Fraud models using clients’ feedback data ensuring these are not only the most accurate, but also the most inclusive.”
Socure is now providing digital onboarding services for more than 750 organizations, and has reported that bookings are up 500 percent in its latest quarterly report. Technology providers like AnyVision and Onfido have also emphasised the need for unbiased AI systems.
October 28, 2021 – by Eric Weiss