To evaluate public perception of the benefits and risks of U.S. Department of Homeland Security use of artificial intelligence technologies, researchers surveyed the nationally representative RAND American Life Panel about the department's uses of these technologies, with a focus on four types of technologies: facial recognition technology, license plate–reader technology, risk-assessment technology, and mobile phone location data.

Public Perceptions of Artificial Intelligence for Homeland Security

Benjamin Boudreaux, Douglas Yeung, Rachel Steratore, Thomas Goode, Nidhi Kalra, Sydne J. Newberry, Michael W. Robbins, Natalia Henriquez Sanchez, Keller Scholl, Victoria M. Smith, et al.

ResearchPublished Mar 20, 2024

Cover: Public Perceptions of Artificial Intelligence for Homeland Security
Download PDF
Additional Downloads

Includes other ebook formats

Artificial intelligence (AI) systems could be crucial in supporting the U.S. Department of Homeland Security's (DHS's) core missions. DHS already uses AI in homeland security missions, and it seeks to further integrate emerging AI capabilities in other applications across DHS components.

However, the full potential of DHS use of emerging AI technologies is subject to several constraints, one of which is how people view government use of those technologies. Public perception of government use of technology is important for several reasons, such as to establish trust in and legitimacy of the government, to facilitate necessary funding and legislative support from Congress, and to foster collaboration with technology companies and operational partners.

Some of these key stakeholders have raised concerns about DHS use of AI technologies, including risks that DHS applications violate privacy and civil liberties, exacerbate inequity, and lack appropriate oversight and other safeguards. These concerns could shape or restrict DHS use of technology, so it is important that DHS understand the extent to which the public agrees with the department's approach to addressing these concerns.

Researchers sought to evaluate public perception of the benefits and risks of DHS use of AI technologies. They developed a survey in 2020 with questions about current and planned DHS use of AI technologies, with a focus on four types of technologies: face recognition technology (FRT), license plate–reader technology, risk-assessment technology, and mobile phone location data. The survey was fielded using the RAND American Life Panel, a nationally representative panel of the American public.

Key Takeaways

  • Significant numbers of respondents might not have formed opinions about government use of FRT.
  • Respondents said that security, accuracy, and privacy were more important than speed or convenience for government use of FRT.
  • Respondents reported agreement that the government's use of FRT had both benefits and risks, but they were likelier to acknowledge risks than benefits.
  • Respondents reported agreement that the government's use of FRT raised multiple types of risk, including misuse, inaccuracy, and bias.
  • Less than one-quarter of the panel reported trusting the government's use of FRT.
  • Safeguards could improve the public's comfort with government use of FRT, but these safeguards' overall effect on comfort levels might be limited.
  • Respondents reported agreement that government must meet certain requirements for using FRT.
  • Public support depends more on the application than on the technology.
  • Respondents reported supporting some government applications of FRT (such as criminal investigations) but not others (such as identifying people at protests or public spaces).
  • Respondents' reported support for risk technologies varied by application and data source.
  • Respondents' reported support for license plate readers and mobile phone location data varied by government application.

Recommendations

  • Proactively engage communities that are uncertain or neutral on government use of AI.
  • Deliberately analyze perceptions of benefits and risks of AI from each stakeholder group’s point of view.
  • Focus on applications and safeguards rather than on the type of AI technology.
  • Consider which data sources underpin AI technologies.
  • Build trust in DHS and the broader U.S. government through continued partnerships.
  • Consider partisanship's impact on DHS use of AI.
  • Use multiple methods to routinely engage key stakeholders.
  • Integrate public perception into technology development and acquisition life cycles.
  • Ensure that public-perception studies are timely.

Topics

Document Details

Citation

Chicago Manual of Style

Boudreaux, Benjamin, Douglas Yeung, Rachel Steratore, Thomas Goode, Nidhi Kalra, Sydne J. Newberry, Michael W. Robbins, Natalia Henriquez Sanchez, Keller Scholl, Victoria M. Smith, and Kristin Warren, Public Perceptions of Artificial Intelligence for Homeland Security. Homeland Security Operational Analysis Center operated by the RAND Corporation, 2024. https://www.rand.org/pubs/research_reports/RRA691-1.html.
BibTeX RIS

Research conducted by

This research was sponsored by the Science and Technology Directorate and conducted in the Management, Technology, and Capabilities Program of the RAND Homeland Security Research Division.

This publication is part of the RAND research report series. Research reports present research findings and objective analysis that address the challenges facing the public and private sectors. All RAND research reports undergo rigorous peer review to ensure high standards for research quality and objectivity.

RAND is a nonprofit institution that helps improve policy and decisionmaking through research and analysis. RAND's publications do not necessarily reflect the opinions of its research clients and sponsors.