Trust, AI, and How the Future Caucus is Shaping Policy for the Coming Years

With leadership from members of the Congressional Future Caucus (CFC) and other members of Congress from both parties, the federal government is working toward a more comprehensive strategy surrounding Artificial Intelligence (AI). By taking steps to create and institute a framework of standards for AI, lawmakers are building a foundation for more trustworthy and transparent applications of the technology, as well as attempting to solidify American leadership in the field. CFC Co-Chair Rep. Anthony Gonzalez (R-OH-16) has been particularly active in the debate, writing a letter on the subject to the National Institute of Standards and Technology (NIST) alongside fellow CFC member Rep. Darren Soto (D-FL-9) and other Representatives in December of 2019. This letter requests that NIST establish a specific framework of rules to settle ethical debates, account for societal impacts, and mitigate potentially negative outcomes of the use of AI.

Elected officials and experts alike are worried that the United States could fall behind peer competitors on the international scale. This development would have long-lasting repercussions. The country that first develops and refines comprehensive AI guidelines will have a significant advantage in that they will be able to impose their own value system onto the technology, impacting implementation and potentially leveraging the technology for political influence abroad.

Additionally, lawmakers are eager to address a prominent issue with the technology: the potential for bias and discrimination. There is a concern that AI programs could perpetuate existing discrimination against already vulnerable groups like ethnic minorities or low-income individuals. This would occur through an overreliance on data that includes previous bias, like rejecting applicants of color for housing loans in certain areas because the data set used to inform the program draws on the historical practice of redlining. By requesting that NIST accelerate their timeline to establish guidelines around AI — specifically with regard to issues of bias and discrimination — lawmakers hope to stay ahead of the competition, while instilling a greater sense of trust among the public and ensuring that the technology will not have disproportionately negative consequences for certain communities.

Concern around the emergence and potential application of AI is not limited to lawmakers. According to a recent Brookings report, 84% of Americans are worried about AI technology to the degree that they believe it should be tightly managed and regulated. In the hope of addressing these fears, as well as sussing out the prominent problems in the field, the Senate Committee on Commerce, Science, and Transportation recently held a hearing that convened experts and leaders in developing the federal government’s AI strategy. NIST Director Dr. Walter Copan briefed senators on the work currently underway at the Institute to ensure explainability and trustworthiness in AI, emphasizing his organization’s goal of inspiring confidence through “the study and deployment of AI methods, tools, and standards”. At the same hearing, France Córdova, the Director of the National Science Foundation (NSF), weighed in to suggest that education was another key aspect of increasing trust in AI through creating a broader knowledge base in the public sphere. By educating a workforce that is prepared to handle the technological advances of the future, Córdova stated that the United States can become — and most importantly remain — a leader in the field.

Turning to specific threats manifested through the use of AI, senators at the January 15th hearing asked a number of questions about “deepfakes,” referring to a method of manipulating photographs and videos through AI technology, and one of the most immediate concerns about the abuse of AI. While many are worried about this application for their own personal privacy, senators emphasized the risk with regard to national security. Senator Maria Cantwell (D-WA), Ranking Member of the Committee, asked a number of questions to witnesses about the technology and seemed especially focused on the ways in which deepfakes could be used to incite an international incident through the mass dissemination of misinformation. Both Córdova and Copan responded that the United States most of all needs to be a leader in the field in order to enforce our ethical and moral standards around its development. Córdova pointed out that the National Science Foundation in particular is working within both the public and private sectors to prevent the misuse of AI by striving to make sure that any utilization, including deepfakes, is as transparent, honest, and fair as possible.

In the fall of last year, Rep. Gonzalez once again took the lead in Congress by introducing H.R. 4355, a bipartisan bill cosponsored by multiple other members of the Congressional Future Caucus. The bill instructed the Director of the National Science Foundation to support new research into deepfakes and other related technologies that could become prevalent in the future. The bill passed the House of Representatives by voice vote on December 9th, 2019 and has been referred to the Senate Commerce, Science, and Transportation Committee. Through the introduction of this bill, supporters hope to mitigate vulnerability and ensure that the United States does not lag behind the rest of the world in AI technology. The clearest steps that the federal government is taking at the moment are to increase support for education in the field and promote research for standards surrounding AI. Both of these actions are targeted at making AI more trustworthy and less intimidating for the American public, as well as ensuring that the United States is not left without influence in the emerging world of Artificial Intelligence.

By: Adam Isaacs-Falbel