New York’s Department of Financial Services is looking into the accusations of sexism, which began with a tweet from David H. Hansson, a software developer who is known for creating Ruby on Rails. Amongst other things, he said the “sexist program” gave him 20 times the credit limit of his wife, who shares equally in all his assets and has a better credit score than him. Hansson referred to the algorithm that determined these credit limits as “a black box” and shared his frustration that the Apple representatives continue to blame the algorithm without sharing any greater reason why the algorithm should have made such a decision. Following Hansson’s tweets, Apple co-founder Steve Wozniak shared on Twitter that he and his wife had the same problem. In this instance Wozniak had 10 times the credit limit of his wife, rather than 20.
While the Apple representatives explained to Hansson that the discrepancy was caused by “just the algorithm,” the laws in New York might have a different perspective. The New York DFS stated that “[a]ny algorithm that intentionally or not results in discriminatory treatment of women or any other protected class violates New York law.” The investigation is currently focused on Goldman Sachs, who is responsible for making the credit decisions for the cards, rather than on Apple.
A Goldman Sachs spokesman claimed that “[i]n all cases, we have not and will not make decisions based on factors like gender,” and explained that family members are evaluated independently using factors like income, amount of debt, and creditworthiness. After the Twitter comments gained media attention, Goldman Sachs changed Mrs. Hansson’s credit limit to match her husband’s without asking for further documentation. Goldman Sachs did not disclose how these factors resulted in such a massive difference in credit limit between a husband and wife.
Many instances of bias in algorithms are currently being investigated across the tech industry. An allegation that UnitedHealth Group has a racist algorithm is also under investigation by the DFS, and multiple instances of bias in other credit algorithms are being investigated by Congress. In 2017, Amazon abandoned an AI tool that they were using to rate job candidates when it was discovered the program unintentionally discriminated based on gender. In that instance, the program learned the discriminatory tendency from a data set of previous job applicants.
The investigation remains at the state level for now, though some voices in Washington, including presidential candidate Sen. Elizabeth Warren (D-Mass.), are calling for a federal response.