Home PublicationsCommentary Survey: Few Americans Want Government to Limit Use of Facial Recognition Technology, Particularly for Public Safety or Airport Screening

Survey: Few Americans Want Government to Limit Use of Facial Recognition Technology, Particularly for Public Safety or Airport Screening

by Daniel Castro
Customer uses facial recognition as identification at TSA security checkpoint

Only one in four Americans (26 percent) think government should strictly limit the use of facial recognition technology, according to a new survey from the Center for Data Innovation—and that support drops even further if it would come at the expense of public safety. Fewer than one in five Americans (18 percent) would agree with strictly limiting the technology if that is the tradeoff, while a solid majority (55 percent) would disagree.

Similarly, only 20 percent of Americans say government should strictly limit use of facial recognition if it would mean airports can’t use the technology to speed up security lines, while a 54 percent majority would disagree with such a limit. And just 24 percent want strict limits if it would prevent stores from using the technology to stop shoplifting, while 49 percent would oppose such a tradeoff.

Table 1: U.S. Internet users’ opinions on facial recognition technology.

There were some differences in these opinions based on age, with older Americans more likely to oppose government limits on the technology. For example, 52 percent of 18 to 34-year-olds opposed limitations that come at the expense of public safety, compared to 61 percent of respondents ages 55 and older. In addition, women were less likely to support limits than men. For example, only 14 percent of women support strictly limiting facial recognition if it comes at the expense of public safety, versus 23 percent of men.

Table 2: U.S. Internet users’ opinions on facial recognition technology, by age and gender.

The survey asked whether police should be allowed to use facial recognition to help find suspects. Support for using the technology that way increases depending on its accuracy: If the software is right 80 percent of the time, then 39 percent agree with using it and 32 percent disagree. If the software is right 90 percent of the time, then 47 percent of respondents agree with using it and 25 percent disagree. And if the software is right 100 percent of the time, then 59 percent agree with using it, while 16 percent disagree.

Table 3: U.S. Internet users’ opinions on use of facial recognition technology by police.

The survey also asked respondents whether government should limit surveillance cameras, since they are integral to many applications of facial recognition technology. Overall, Americans were more likely to support limiting surveillance cameras (36 percent) than facial recognition technology (26 percent). But that flips when respondents are asked about tradeoffs. For example, if it would mean stores couldn’t use the technology to stop shoplifting, then support for limits on surveillance cameras drops by half, from 36 percent to just 18 percent, while support for limits on facial recognition slips only slightly from 26 percent to 24 percent.

If it would come at the expense of public safety, then just 18 percent of Americans would agree with limiting surveillance cameras and the same percentage would agree for facial recognition. These findings suggest that what little support there is for limiting facial recognition technology is related to existing support for limiting the use of surveillance cameras.

Table 4: U.S. Internet users’ opinions on regulating surveillance cameras and facial recognition technology.

Survey Questions

1. Agree or disagree? The government should strictly limit the use of surveillance cameras.

  • Strongly agree: 16.1 percent
  • Somewhat agree: 20.1 percent
  • Neither agree nor disagree: 34.4 percent
  • Somewhat disagree: 14.2 percent
  • Strongly disagree: 15.2 percent

2. Agree or disagree? The government should strictly limit the use of surveillance cameras even if it means stores can’t use them to reduce shoplifting.

  • Strongly agree: 7.4 percent
  • Somewhat agree: 10.8 percent
  • Neither agree nor disagree: 23.0 percent
  • Somewhat disagree: 22.5 percent
  • Strongly disagree: 36.3 percent

3. Agree or disagree? The government should strictly limit the use of surveillance cameras even if it comes at the expense of public safety.

  • Strongly agree: 7.4 percent
  • Somewhat agree: 10.5 percent
  • Neither agree nor disagree: 23.5 percent
  • Somewhat disagree: 21.5 percent
  • Strongly disagree: 37.1 percent

4. Agree or disagree? The government should strictly limit the use of facial recognition technology.

  • Strongly agree: 12.5 percent
  • Somewhat agree: 13.7 percent
  • Neither agree nor disagree: 29.0 percent
  • Somewhat disagree: 20.5 percent
  • Strongly disagree: 24.4 percent

5. Agree or disagree? The government should strictly limit the use of facial recognition technology even if it means stores can’t use it to reduce shoplifting.

  • Strongly agree: 12.2 percent
  • Somewhat agree: 11.6 percent
  • Neither agree nor disagree: 27.1 percent
  • Somewhat disagree: 20.9 percent
  • Strongly disagree: 28.2 percent

6. Agree or disagree? The government should strictly limit the use of facial recognition technology even if it means airports can’t use it to speed up security lines.

  • Strongly agree: 9.6 percent
  • Somewhat agree: 10.3 percent
  • Neither agree nor disagree: 25.7 percent
  • Somewhat disagree: 20.5 percent
  • Strongly disagree: 33.8 percent

7. Agree or disagree? The government should strictly limit the use of facial recognition technology even if it comes at the expense of public safety.

  • Strongly agree: 8.2 percent
  • Somewhat agree: 10.1 percent
  • Neither agree nor disagree: 26.9 percent
  • Somewhat disagree: 21.1 percent
  • Strongly disagree: 33.8 percent

8. Agree or disagree? Police departments should be allowed to use facial recognition technology to help find suspects if the software is correct 80% of the time.

  • Strongly agree: 17.7 percent
  • Somewhat agree: 21.6 percent
  • Neither agree nor disagree: 28.5 percent
  • Somewhat disagree: 15.2 percent
  • Strongly disagree: 16.9 percent

9. Agree or disagree? Police departments should be allowed to use facial recognition technology to help find suspects if the software is correct 90% of the time.

  • Strongly agree: 23.4 percent
  • Somewhat agree: 23.9 percent
  • Neither agree nor disagree: 27.7 percent
  • Somewhat disagree: 11.3 percent
  • Strongly disagree: 13.7 percent

10. Agree or disagree? Police departments should be allowed to use facial recognition technology to help find suspects if the software is correct 100% of the time.

  • Strongly agree: 41.1 percent
  • Somewhat agree: 18.3 percent
  • Neither agree nor disagree: 24.5 percent
  • Somewhat disagree: 6.0 percent
  • Strongly disagree: 10.0 percent

Age:

  • 3.4% 18-24, 12.3% 25-34, 17.6% 35-44, 22.5% 45-54, 23.3% 55-64, 21.0% 65+ in sample
  • 12.6% 18-24, 19.4% 25-34, 17.5% 35-44, 17.6% 45-54, 16.7% 55-64, 16.2% 65+ in target population

Gender:

  • 56.4% male. 43.6 percent female in sample
  • 48.1% male, 51.9% female in target population

Region:

  • 16.0% Northeast, 34.0% South, 28.5% Midwest, 21.5% West in sample
  • 17.3% Northeast, 37.5% South, 21.3% Midwest, 23.9% West in target

Detailed Survey Results

Download additional detail about the survey results.

Survey Methodology

The Center for Data Innovation conducted a national online poll of 3,151 U.S. adult Internet users between December 13, 2018 and December 16, 2018. Using Google Surveys, we applied weights to each response to match the breakdowns of age, gender, and region to those demographic breakdowns in the national Internet population as estimated by the U.S. Census Bureau’s 2015 Current Population Survey (CPS) Computer and Internet Use Supplement.

Multiple analyses have found Google Surveys to be a useful survey tool. In 2012, the Pew Research Center compared the results for 43 questions it asked through telephone surveys and Google Surveys, finding that the median difference between the two methods’ results was three percentage points. Moreover, Google Surveys accurately predicted the 2012 presidential election. Lastly, a 2016 analysis, published in the peer-reviewed journal Political Analysis by Rice University political scientists, replicated four canonical social science experiments with Google Surveys and concluded that Google Surveys “is likely to be a useful platform for survey experimenters doing rigorous social scientific work.”

Google Surveys donated the use of its platform for this research but played no role in the findings or in developing the questions. To learn more about Google Surveys’ methodology and accuracy, please see the Google Surveys Whitepaper and a study comparing Google Surveys to other Internet surveys.

About the Center for Data Innovation

The Center for Data Innovation conducts high-quality, independent research and educational activities on the impact of the increased use of information on the economy and society. In addition, the Center for Data Innovation formulates and promotes pragmatic public policies designed to enable data-driven innovation in the public and private sector, create new economic opportunities, and improve quality of life. The Center is a nonprofit, nonpartisan research institute affiliated with the Information Technology and Innovation Foundation.

Image credit: Delta News Hub

You may also like

Show Buttons
Hide Buttons