With Vice President Kamala Harris on the Democratic ticket for the 2024 presidential election, experts are weighing what A.I. policy could look like under a Harris administration. Many foresee Harris continuing a path of measured yet fairly aggressive steps to regulate A.I., but the president hopeful may have more on her mind.
“There’s a lot of track record here,” Steve Wilson, a project leader for the Open Web Application Security Project and chief product officer of the cybersecurity company Exabeam, told Observer, referencing Harris’s involvement in the Biden administration’s A.I. Executive Order last year and subsequent action items.
Tech entrepreneurs Marie Roker-Jones and Annie Brown, who serve as two of the co-chairs of Founders for Kamala, a grassroots political organization, see Harris as emboldening technological innovation with diversity, equity and inclusion in mind.
“The Biden-Harris administration has been really good about working towards getting more resources for underrepresented entrepreneurs,” Brown told Observer. Brown, an entrepreneur-in-residence at UC San Diego and the founder of Reliabl, a company making A.I. infrastructure software for businesses, recently attended a White House roundtable to promote more inclusion in the venture capital industry.
“One thing that came up is that more diverse founders need access to capital,” she said, “ because if you don’t have women, people of color, LGBTQ, disabled persons working on these tools and building these tools and getting the funding to build A.I. tools, then you’re going to recreate a lot of bias.”
IBM defines A.I. bias as “systems that produce biased results that reflect and perpetuate human biases within a society, including historical and current social inequality.” Datasets, algorithmic models and A.I. outputs can all elicit bias. But because of Harris’s experience of being falsely accused of sharing doctored rally photos, “it’s more personal for her to understand the risk of how A.I. can be used to further weaponize vulnerable and marginalized communities,” Roker-Jones, a venture capital scout and serial tech founder, told Observer.
Recently, Harris has targeted Big Tech companies for monopolistic behavior. By attacking this issue head-on before A.I. gets any bigger, Brown said, it could foster more competition and innovation down the line. Brown said of side effects of condensed competition in A.I., “There’s less motivation for them to produce good work, but also you get less viewpoints in that A.I., less training, less data sets in that A.I., and then the bias becomes worse and worse.”
Because technologies like A.I. are heavily linked to data, the European Union’s existing regulation on data privacy, the GDPR, as well as its forthcoming A.I. initiatives, serve as a solid foundational model for the U.S. While the U.S. has not created its own cohesive data privacy law at the federal level, Biden’s A.I. executive order has so far generated voluntary commitments from 15 major technology companies (including Apple) to follow frameworks on managing generative A.I. and data security risks as outlined by the National Institute of Standards and Technology (NIST). Meanwhile, certain states, including California, Virginia and Colorado, have already instated their own versions of the GDPR.
“There are definitely ways that you could regulate this that would really blunt innovation,” Wilson said. “Measured, proper guidance will enable people to feel more comfortable with these technologies and adopt them faster—and the big challenge for whoever’s in the next administration is: How do you do that in a measured way without going overboard?”
