Democratic presidential candidate Joe Biden chose Sen. Kamala Harris (D-CA) to be his vice presidential running mate today. Born in Oakland, California and raised in Berkeley, Harris is the first African American woman and first Asian American woman to be chosen as a U.S. vice presidential candidate on a major party ticket. Though a vice presidential pick can sometimes be considered ceremonial, if elected, Biden will become the oldest president in U.S. history at 78 years old.
How Harris stands on a range of issues is going to be meticulously analyzed in the days leading up to the Democratic National Convention and the weeks leading up to Election Day. A former prosecutor, Harris is known as a member of the Senate Judiciary committee, but she’s also led cybersecurity proposals in Congress, and raised tech policy issues where she grew to prominence in the San Francisco Bay Area. Here we look back at how Kamala Harris the U.S. Senator and presidential candidate reacted to issues at the intersection of AI and policy.
In fall 2018, Harris joined other members of the U.S. Senate in sending a series of letters to federal agencies aimed at addressing algorithmic bias based on race, gender, or other characteristics.
One letter asked the Equal Employment Opportunity Commission (EEOC) how it investigates claims of algorithmic bias in hiring practices and whether the agency considers use of facial recognition a violation of existing civil rights and workplace anti-discrimination law.
The FTC letter imagines a scenario in which a Black woman is falsely arrested due to the use of facial recognition software. Some of the first known cases of false arrests of Black men due to misidentification with facial recognition took place in June.
The FBI letter demanded to know whether the FBI had responded to a 2016 request by the Government Accountability Office (GAO) report recommending the FBI test its facial recognition and take other steps to ensure its accuracy. The facial recognition system is used by the FBI as well as state and local officials. Months later, the FBI’s failure to act on the GAO’s recommendation became a main subject of discussion in a House Oversight and Reform committee hearing.
Each of the letters cite an ACLU study that misidentified several members of Congress as criminals and the work of Joy Buolamwini and the Gender Shades project that facial recognition systems from companies like IBM and Microsoft performed better for white men than they do for women with dark skin.
As a presidential candidate, Harris introduced a criminal justice plan that said she would work with both civil rights groups and law enforcement to ensure facial recognition and other technology do not advance racial bias, and invest federal funding to convince state and local officials to do the same. Other candidates like Bernie Sanders supported an outright ban of facial recognition use by federal law enforcement like the FBI.
Her stance on issues like predictive policing and facial recognition may be especially important given fervent calls for police reform and the end of institutional racism rose to international prominence in June. They may also be important given Harris’ past record as California Attorney General, San Francisco District Attorney, and a public prosecutor in Alameda County.
Federal AI policy
Harris was one of four U.S. Senators behind the AI in Government Act, a bill first introduced in 2018 and then reintroduced last year that would have helped develop a more cohesive federal AI policy through the General Services Administration (GSA).
The AI in Government Act of 2019 would direct the heads of each federal agency to recommend ways to remove barriers to AI adoption as well as “best practices for identifying, assessing, and mitigating any discriminatory impact or bias on the basis of any classification protected under Federal nondiscrimination laws, or any unintended consequence of the use of artificial intelligence by the Federal Government.”
It would also create a Center of Excellence within the General Services Administration to assist federal agencies in acquiring AI services and supply technical expertise. The Center of Excellence would also advise the White House Office of Science of Technology and consult with the Pentagon as well as the National Science Foundation and help develop policy related to the use of AI by federal agencies.
A unified AI policy that considers the nation’s research and development needs and broader strategy is considered important for economic activity but is increasingly being tied to research and development strategies of national governments and militaries. In the U.S. and China, governments have set goals to maintain a kind of AI supremacy over other nations.
In the end, the AI in Government Act, like a range of bills introduced in recent years to regulate AI and the formation of a robust U.S. strategy, has yet to come up for a vote, but it was unique in that the bill attracted bipartisan support. Harris wasn’t the earliest lawmaker to introduce AI regulation nor is she the most associated with facial recognition regulation today, but the series of letters and legislation give you an idea of where Harris stands on AI. So does her repeated expressed concern with the potential for AI to perpetuate bias.
“When we look also at these emerging fields and look at issues like AI — artificial intelligence and machine learning — there is a real need to be very concerned about how being built into it is racial bias,” Harris said in a video posted by her Senate office in April 2019. “It’s a real issue and it’s happening in real time, and the thing about racial bias in technology is that unlike the racial bias that you can pretty easily, all of us can detect when you get stopped in a department store or while you’re driving, the bias that is built into technology will not be very easy to detect, and so machine learning is literally that the machine is learning, and it’s learning what it’s being fed, which is going to be a function of who’s feeding it and what it’s being fed.”
The Biden-Harris 2020 campaign and potential presidency may take stances different than Kamala Harris the Senator or presidential candidate. Nonetheless, Harris has introduced early legislation aimed at creating a cohesive AI strategy for the federal government and has a seemingly open disdain for tech with the potential, and proven likelihood, to perpetuate bias.
Kamala Harris can be called the first person on a major party ticket in a lot of categories, but however you consider it, Harris appears to be one of the most informed candidates on a major party’s presidential ticket to take national AI policy and algorithmic bias seriously.