May 1, 2018 | Number 5

ImmigrationR E S E A R C H A N D P O L I C Y B R I E F

Drones on the BorderEfficacy and Privacy ImplicationsBy David J. Bier and Matthew Feeney

In response to President Donald Trump’s call for a border wall, some members of Congress have instead offered a “virtual wall”—ocean-to-ocean border surveillance with technology, especially unmanned aircraft known as drones. U.S. Cus-

toms and Border Protection (CBP) already operates a fleet of nine unmanned aircraft. Although drones have been widely used in foreign battlefields, they have failed to help CBP apprehend illegal border crossers and seize drugs. Drones have led to only 0.5 ercent of apprehensions at a cost of $32,000 per arrest.

At the same time, drones undermine Americans’ privacy. Their surveillance records the daily lives of Americans living along the border, and because CBP regularly uses its drones to support the operations of other federal agencies as well as state and local police, its drones allow for government sur-veillance nationwide with minimal oversight and without warrants. CBP should wind down its drone program and, in the meantime, establish more robust privacy protections.

BACKGROUNDSince fiscal year 2006, CBP’s Air and Marine Operations has

operated Predator B drones, a version of the military’s MQ-9 Reaper drone, along the border.1 A drone crew consisting of a pilot, a sensor operator, and a radar operator controls the air-craft and relays information about suspected crossings to the U.S. Border Patrol.2 In September 2017, CBP began testing

smaller hand-launched drones, including AeroVironment’s Raven and Puma small unmanned aircraft systems.3 In the summer of 2016, the agency solicited offers for small drones with facial recognition capability.4 The agency will finish its review of the hand-launched systems in spring 2018.5

LIMITATIONS OF DRONESThe biggest difficulty with drones is positioning them

to surveil the border. Although CBP often comes close to or exceeds its flight-hour goals, its Predators were airborne only 6.4 percent of available hours per year from 2013 to 2016 (Table 1)—on average just 1 hour and 35 minutes per day.6 CBP’s other aerial assets, such as unmanned buoyant aircraft outfit-ted with Tethered Aerostat Radar System (TARS), conduct surveillance at 15,000 feet and operate around 60 percent of the time.7 A Predator B drone engine cannot regularly handle its maximum flight of 20 hours; even 6 or 7 hours of operation per day would cut at least 15 years off the 20-year lifespans of such engines.8

CBP managed to conduct only 69 percent of its scheduled drone launches in 2016.9 The agency canceled 20 percent of its drone flights because of weather alone. When the flights do take place, weather can prevent drones from flying in the highest value airspace.10 Even occasional cloud cover can disrupt surveillance. In 2010, CBP allowed CNN to record an incident where operators tracked a group across the bor-der, but just as the pilots started to instruct the agents on the

David J. Bier is an immigration policy analyst at the Cato Institute’s Center for Global Liberty and Prosperity. Matthew Feeney is a policy analyst at the Cato Institute.

2

ground, clouds blocked their view, and the subjects escaped.11 Unlike U.S. Border Patrol agents in the field, drones cannot apprehend subjects.

Moreover, two of the CBP drones (18 percent of its fleet) have crashed within just 10 years of purchase—one because of human error in 2006 and another owing to a generator failure in 2014.12 The U.S. Air Force has crashed the same share of its version of the Predator B.13 Of the 15 CBP or military crashes where the cause is publicly known, seven occurred because of mechanical failure, five were caused by pilot error, two by a combination of mechanical and pilot failure, and one by the weather.14

CBP drones also lack the authority to fly in certain areas. They require advanced permission from the U.S. Depart-ment of Defense to fly in restricted airspace and they cannot

fly in areas with a high volume of commercial flights.15 Such restrictions block off large areas near San Diego, California, and Yuma, Arizona. Drones regularly patrolled only about 170 miles of the 2,000-mile southern land border in 2013, according to the most recent available data.16 For these rea-sons, CBP drones fail to replace or displace any other surveil-lance equipment or agents in the field.

EFFECTIVENESS OF DRONESWhen CBP drones do fly, they contribute little to border

security. From 2013 to 2016, the U.S. Border Patrol attrib-uted fewer than 8,000 of its 1.7 million apprehensions to drones (Table 2).17 During the same years, only 3 percent of all U.S. Border Patrol marijuana seizures—which account

Table 1Total flight hours for CBP Predator B drones and share of total available hours, 2013–2016

  2013 2014 2015 2016 Total

Total flight hours 5,110 4,615 5,515 5,540 20,780

Share of total available hours (percent) 5.8 5.6 7.0 7.0 6.4

Source: Government Accountability Office.

Table 2Apprehensions attributable to drones and all apprehensions, 2013—2016

  2013 2014 2015 2016 Total

Drone apprehensions 2,252 1,834 2,136 1,729 7,951

All apprehensions 420,789 486,651 337,117 415,816 1,660,373

Drone share (percent) 0.5 0.4 0.6 0.4 0.5

Sources: United States Border Patrol; Government Accountability Office.

Table 3Marijuana seizures attributable to drones and all seizures, 2013—2016

  2013 2014 2015 2016 Total

Drone seizures 56,308 68,624 68,913 29,972 223,817

All seizures 2,430,123 1,922,545 1,538,307 1,294,052 7,185,027

Drone share (percent) 2.3 3.6 4.5 2.3 3.1

Sources: United States Border Patrol; Government Accountability Office.

3

for 99.3 percent of Border Patrol’s drug seizures by weight—occurred in part because of drones (Table 3).18 The agency has not provided any estimate of how many of these seizures or apprehensions would have occurred anyway had drones not identified them.

CBP asserts that drones detect patterns of illegal entries, allowing the agency to reposition its assets accordingly. However, only 2 percent of drone missions resulted in any evidence of a previously undetected illegal crossing in 2014.19 Moreover, the Government Accountability Office found that drones detected only 21,384 suspected border crossers from 2014 to 2016.20 Yet during this time, the U.S. Border Patrol apprehended 1.2 million people while an estimated 570,000 evaded capture.21 In other words, drones may have detected at most 1.2 percent of total illegal border crossings.

COST OF DRONESEach Predator B drone costs $17 million to purchase and

$12,255 per flight hour to operate.22 Thus, CBP’s drone pro-gram cost a grand total of $255 million from 2013 to 2016 (Table 4).23 These figures likely understate the cost of the systems’ depreciation because they assume a 20-year lifes-pan, but 18 percent of CBP drones crashed in their first 10 years. For comparison, manned aircraft with surveillance capabilities similar to the Predator B cost only about $1,500 to $2,000 per flight hour.24 Each drone apprehension costs the federal government $32,000 (Table 5). This cost of drone

apprehension compares with the average cost of apprehen-sion of less than $9,000.25

In 2014, the Office of Inspector General for the Depart-ment of Homeland Security (DHS) found that CBP’s Preda-tor B drone program’s effectiveness was unproven and did not meet program expectations.26 In 2016, DHS recom-mended that CBP not expand its drone program because of operational difficulties and exorbitant costs.27

PRIVACY IMPLICATIONSNo law requires CBP to obtain a warrant before using

drones for surveillance, and the U.S. Supreme Court has upheld warrantless aerial surveillance in three cases.28 The CBP may establish its standards for surveillance and conduct operations without a court order or even suspicion that those under surveillance have committed a crime, thereby allow-ing the CBP to freely use its surveillance authority to collect information on the lives of law-abiding U.S. residents inside the United States. The quality of the drone footage and the long distances from the physical border where drones often operate considerably amplify this concern. Congress should establish clear and strict rules for CBP drone operations.

Drones allow for real-time, high-quality video feeds of the ground.29 Predator B drones are supposed to avoid urban areas, and CBP states that it does not deploy them to monitor protests and other activities protected by the First Amendment.30 Nonetheless, they collect enormous amounts

Table 4Flight hours and costs for Customs and Border Protection Predator B drones, 2013–2016

  2013 2014 2015 2016 Total

Flight hours 5,110 4,615 5,515 5,540 20,780

Total costs $62,623,050 $56,556,825 $67,586,325 $67,892,700 $254,658,900

Sources: Department of Homeland Security Office of Inspector General; Government Accountability Office.

Table 5Cost per apprehension by source of apprehension—Predator B drones and overall, 2013–2016

  2013 2014 2015 2016 Average

Overall $8,239.00 $7,469.12 $11,265.59 $8,760.65 $8,758.50

Drones $27,807.75 $30,837.96 $31,641.54 $39,267.03 $32,028.54

Sources: Author’s calculations based on Department of Homeland Security Office of Inspector General; Government Accountability Office; Customs and Border Protection.

4

of information on U.S. residents, including citizens, in bor-der regions. CBP’s Federal Aviation Administration (FAA) drone authorization allows CBP drones to operate along and within 25 to 60 miles of the southern border and along and within 100 miles of the northern border.31 Residents are well aware of drone surveillance and may change their behavior in response.32

Each Predator B carries a ball camera that can stream live video in electrical optimal (Images 1 and 2) or infrared (Image 3).33 Some CBP drones also use Synthetic Aperture Radar, which allows for digital comparison of two areas of the border at different times to detect patterns in crossings.34 Two CBP drones in Arizona can carry VADER (Vehicle and Dismount Exploitation Radar), a brand of airborne tactical radar.35 VADER allows the drone to identify moving objects such as persons or vehicles and track them on the ground in real-time.

The FAA requires Predator drones to fly between 19,000 to 28,000 feet.36 According to DHS, images from a Predator drone’s current camera cannot identify an individual’s exact height, weight, eye color, hairstyle, or facial image from such a high altitude, but they can identify backpacks, vehicle types, and firearms (Image 2).37 These limitations would not apply to the smaller drones that CBP is currently testing.38 Small drones fly closer to the ground and can identify people’s eth-nicity, height, weight, and hairstyle. Image 4 is a frame from a video from one of CBP’s small drones.39 Small drones can identify faces, and operators are able to read license plates, signs, banners, and other writing from the air.

Worryingly, CBP wants to combine its small drones with facial recognition software that could compare drone- captured images to a vast array of images in law enforcement databases.40 The FBI’s Facial Analysis, Comparison, and Evaluation (FACE) Services unit allows the agency to access driver’s license photos from 16 states and 411 million images from other sources, including the State and Defense depart-ments.41 The FBI’s Next Generation Identification and the Interstate Photo System (NGI-IPS) include more than 30 million photos.42 Around half of all American adults are in a law enforcement facial recognition network.43

Facial recognition programs are prone to errors.44 A May 2016 Government Accountability Office report revealed that, although the FBI did not rigorously test the accuracy of NGI-IPS prior to deployment, what testing did take place found that the software failed 14 percent of the time to include the correct person in a 50-person list.45 Research on a num-ber of facial recognition systems has found that their error rates are not evenly distributed among races and genders.46 If

CBP’s small drones begin to use facial recognition technology, law enforcement agencies run the increased risk of detaining law-abiding people as suspected border crossers.

NON-BORDER PATROL USES OF BORDER DRONES

CBP often uses its aircraft to assist non-Border Patrol operations.47 From 2010 to 2012, CBP operators flew 687 missions on behalf of other agencies.48 From 2013 to 2016, only about half of CBP drone flight hours were actually in support of Border Patrol.49 Furthermore, CBP reports that 20 percent of all Predator B flights were not in coastal or border areas.50 The cooperation between CBP and other federal agencies means that Americans living near the border aren’t the only ones who risk having their privacy infringed by CBP drones.

State and local agencies also often request CBP drone assistance for routine law enforcement matters. From 2013 to 2016, CBP drones flew 416 flight hours for state and local police, often failing to record which police departments were requesting CBP drone assistance.51 The first reported case of a drone aiding a domestic arrest occurred in 2011 when CBP deployed a drone without a warrant in North Dakota to determine whether suspects that local police were seeking to arrest were armed.52 A district judge rejected one of the sus-pect’s motion to suppress the warrantless drone surveillance, writing that “there was no improper use of an unmanned aerial vehicle.”53

DHS does not indicate whether CBP adheres to state warrant requirements when assisting local law enforcement agencies.54 For instance, Texas requires local police to obtain warrants to conduct searches with drones, whereas the other three southwest border states do not.55 Thus, CBP drones could be a loophole allowing local police to evade local demo-cratic scrutiny of drone use or the requirement for a search warrant. Unfortunately, CBP fails to release sufficient data on local support operations to identify exactly how they are used.

RECOMMENDATIONSIf CBP chooses to continue flying expensive and ineffec-

tive drones, the agency should at least mandate the following privacy protections:

� CBP should use its drones solely for border security operations except in the case of states of emergency.

5

� CBP should not conduct drone surveillance more than five miles from the border.

� If CBP does use its drones to support state and local operations, it should ensure that its drone pilots com-ply with state and local drone legislation, including warrant requirements.

� CBP should not seek drones with facial recognition capability, which puts law-abiding Americans’ privacy at increased risk.

� At least six months before deploying new surveillance technology, CBP should disclose details about the technology’s capabilities, including information about the type of data to be collected, how long CBP plans to keep the data, when CBP will share the data, and with whom it will share the data.

� CBP should study replacing drones with surveillance

technology that limits unnecessary data collection on U.S. residents.

CONCLUSION CBP’s drone program has failed to live up to its expecta-

tions. Its expense, disproportionately small contribution to border security, and infringement on Americans’ privacy are good reasons for CBP to wind down its drone program. How-ever, if CBP does continue to use drones, it should further constrain their use to prevent unnecessary data collection of Americans. China and other authoritarian countries have already begun to use drones to conduct domestic suspicion-less surveillance, including of protesters and dissidents.56 The United States should put in place safeguards to prevent similar actions here.

Image 2Men with rifles in Iraq

Source: L’Espresso.

Image 1CBP Predator electrical optimal

Source: U.S. Department of Homeland Security.

Image 3Predator infrared image

Source: CNN.

Image 4CBP small drone image of the border

Source: U.S. Customs and Border Protection.

6

NOTES1. Government Accountability Office, “Additional Actions Need-ed to Strengthen Collection of Unmanned Aerial Systems and Aerostats Data,” GAO-17-152, February 2017, http://www.gao.gov/assets/690/682842.pdf; and Alberto Cuadra and Craig Whitlock, “How Drones Are Controlled,” Washington Post, June 20, 2014, http://www.washingtonpost.com/wp-srv/special/national/drone-crashes/how-drones-work/.

2. Office of Inspector General, “CBP’s Use of Unmanned Aircraft Systems in the Nation’s Border Security,” U.S. Department of Homeland Security, OIG-12-85, May 2012, https://www.oig.dhs.gov/assets/Mgmt/2012/OIG_12-85_May12.pdf.

3. “CBP to Test the Operational Use of Small Unmanned Air-craft Systems in 3 U.S. Border Patrol Sectors,” CBP Media Re-lease, September 14, 2017, https://www.cbp.gov/newsroom/national-media-release/cbp-test-operational-use-small-unmanned-aircraft-systems-3-us-border; Avinc, “Puma™ AE RQ-20B,” 2017, https://www.avinc.com/images/uploads/product_docs/PumaAE_Datasheet_2017_Web_v1.1.pdf; and Avinc, “Ra-ven® RQ-11A/B,” 2017, https://www.avinc.com/images/uploads/product_docs/Raven_Datasheet_2017_Web_v1.pdf.

4. FedBizOpps, “Small Unmanned Aircraft System (sUAS): Solici-tation Number: HSHQDC-16-R-00114,” Department of Home-land Security, July 15, 2016, https://www.fbo.gov/index?s=opportunity&mode=form&id=5bb697a0dd83dccb4e011e905865f914& tab=core&_cview=0.

5. Mark Rockwell, “Border Agencies Test Hand-Launched Drones,” FCW, September 15, 2017, https://fcw.com/articles/ 2017/09/15/border-patrol-drone-tests.aspx.

6. This represents 5,540 combined flight hours/(365 days x 24 hours x 9 drones): Government Accountability Office, “Addition-al Actions Needed to Strengthen.”

7. Government Accountability Office, “Additional Actions Need-ed to Strengthen.”

8. Customs and Border Protection, “CBP Responds to Inspec-tor General Report on Unmanned Aircraft Program,” Septem-ber 28, 2016, https://www.cbp.gov/newsroom/video-gallery/video -library/cbp-responds-inspector-general-report-unmanned-aircraft-program.

9. Government Accountability Office, “Additional Actions Need-ed to Strengthen,” p. 28.

10. Government Accountability Office, “Additional Actions Needed to Strengthen Collection of Unmanned Aerial Systems and Aerostats Data” p. 49.

11. “Drones Patrol the Border for Illegal Immigrants,” CNN, May 12, 2010, https://youtu.be/yZHILC3vdKI?t=1m48s.

12. Shawn Musgrave, “The Mystery of US Customs’ Crashed Drone,” Motherboard, April 24, 2014, https://motherboard.vice.com/en_us/article/jp5pg4/the-mystery-of-us-customs-crashed-drone; and Arthur Rotstein, “Operator Fired in ’06 Crash of Unmanned Aerial Drone,” Associated Press, October 18, 2007, http://tucson.com/news/local/border/operator-fired-in-crash-of-unmanned-aerial-drone/article_b4cd0e8a-596d-5c5a-bea1-3b7ab1188325.html.

13. Emily Chow, Alberto Cuadra, and Craig Whitlock, “Fallen from the Skies,” Washington Post, June 20, 2014, http://www.washingtonpost.com/wp-srv/special/national/drone-crashes/database/; “2016 USAF Almanac,” Air Force Magazine 99, no. 5 (May 2016), http://secure.afa.org/joinafa/AFMag0516/files/downloads/attachments/0516fullissue.pdf; and Craig Whitlock, “When Drones Fall from the Sky,” Washington Post, June 20, 2014, http://www.washingtonpost.com/sf/investigative/2014/06/20/when-drones-fall-from-the-sky/?utm_term=.c96f4eb5079b.

14. Whitlock, “When Drones Fall from the Sky.”

15. Government Accountability Office, “Additional Actions Needed to Strengthen,” p. 49.

16. Office of Inspector General, “U.S. Customs and Border Pro-tection’s Unmanned Aircraft System Program Does Not Achieve Intended Results or Recognize All Costs of Operations,” De-partment of Homeland Security, OIG-15-17, December 24, 2014, https://www.oig.dhs.gov/assets/Mgmt/2015/OIG_15-17_Dec14.pdf.

17. Government Accountability Office, “Additional Actions Needed to Strengthen”; and United States Border Patrol, “Total Illegal Alien Apprehensions by Fiscal Year,” https://www.cbp.gov/sites/default/files/assets/documents/2017-Dec/BP%20Total%20Apps%2C%20Mexico%2C%20OTM%20FY2000-FY2017.pdf.

7

18. Government Accountability Office, “Additional Actions Needed to Strengthen”; and Customs and Border Protection, “CBP Enforcement Statistics FY2018,” https://www.cbp.gov/newsroom/stats/cbp-enforcement-statistics.

19. Associated Press, “Half of US-Mexico Border Now Patrolled Only by Drone,” Guardian, November 13, 2014, https://www.the-guardian.com/world/2014/nov/13/half-us-mexico-border-patrolled- drone.

20. Government Accountability Office, “Additional Actions Needed to Strengthen.”

21. For 2014 and 2015, CBP reported that an average of 190,000 per year made it across the border. Tribune News Services, “Barely Half of Illegal Border Crossers Caught, According to Homeland Security Report,” Chicago Tribune, October 7, 2016, http://www.chicagotribune.com/news/nationworld/ct-us-mexico-border-crossing-captures-20161006-story.html.

22. Office of Inspector General, “U.S. Customs and Border Protec-tion’s Unmanned Aircraft System Program Does Not Achieve.”

23. Office of Inspector General, “U.S. Customs and Border Protec-tion’s Unmanned Aircraft System Program Does Not Achieve.”

24. Bob Ortega, “Is Pricey Border Patrol Drone Program Worth the Cost?” Arizona Republic, July 21, 2015, https://www.azcentral.com/story/news/arizona/investigations/2015/06/21/border-patrol-drone-program/28999735/v.

25. Obtained by dividing the entire Border Patrol budget by number of apprehensions: United States Border Patrol, “Enacted Border Patrol Program Budget by Fiscal Year (Dollars in Thousands),” 2017, https://www.cbp.gov/sites/default/files/assets/documents/2017-Dec/BP%20Budget%20History%201990-2017.pdf.

26. Office of Inspector General, “U.S. Customs and Border Pro-tection’s Unmanned Aircraft System Program,” p. 1.

27. Government Accountability Office, “Additional Actions Needed to Strengthen.”

28. Dow Chemical Co. v. United States, 476 U.S. 227 (1986); Califor-nia v. Ciraolo, 476 U.S. 207 (1986); and Florida v. Riley, 488 U.S. 445 (1989).

29. Department of Homeland Security, “Privacy Impact Assess-ment for the Aircraft Systems,” DHS/CBP/PIA-018, Septem-ber 9, 2013, https://www.dhs.gov/sites/default/files/publications/privacy-pia-cbp-aircraft-systems-20130926.pdf.

30. Department of Homeland Security, “Best Practices for Pro-tecting Privacy, Civil Rights & Civil Liberties in Unmanned Air-craft Systems Programs,” U.S. Department of Homeland Security Privacy, Civil Rights & Civil Liberties Unmanned Aircraft Sys-tems Working Group, December 18, 2015, https://www.dhs.gov/sites/default/files/publications/UAS%20Best%20Practices.pdf.

31. Department of Homeland Security, “Privacy Impact Assess-ment for the Aircraft Systems”; and United States v. Ramsey, 431 U.S. 606 (1977).

32. Todd Miller, “War on the Border,” New York Times, August 17, 2013, http://www.nytimes.com/2013/08/18/opinion/sunday/war-on-the-border.html.

33. Department of Homeland Security, “Privacy Impact Assess-ment for the Aircraft Systems.”

34. Office of Inspector General, “U.S. Customs and Border Pro-tection’s Unmanned Aircraft System Program.”

35. Office of Inspector General, “U.S. Customs and Border Pro-tection’s Unmanned Aircraft System Program”; Robert Beckhusen, “Army Works to Ensure Homemade Bombs Don’t Escape the Gaze of ‘Vader,’” Wired, December 7, 2012, https://www.wired.com/2012/12/vader/; and Brian Bennett, “Radar Shows U.S. Bor-der Security Gaps,” Los Angeles Times, April 3, 2013, http://articles.latimes.com/2013/apr/03/nation/la-na-border-radar-20130404.

36. Department of Homeland Security, “Privacy Impact Assess-ment for the Aircraft Systems.”

37. U.S. Customs and Border Protection, “Unmanned Aircraft Sys-tem MQ-9 Predator B,” https://www.cbp.gov/sites/default/files/documents/FS_2015_UAS_FINAL_0.pdf; and Gianluca Di Feo, “Così i droni italiani spiano i covi del Califfo,” L’Espresso, Decem-ber 10, 2015, http://espresso.repubblica.it/plus/articoli/2015/12/10/news/esclusivo-espresso-cosi-droni-italiani-spiano-il-califfo-1.242989?refresh_ce.

38. Rockwell, “Border Agencies Test Hand-Launched Drones.”

8

39. “Drone Footage Shows Prototype Border Walls at U.S.-Mexico Border,” Washington Post, October 19, 2017, https://www.washingtonpost.com/video/national/drone-foota ge- shows -prototy pe-border-wal ls -at-us -mexico-border/2017/10/19/3d79c380-b4b7-11e7-9b93-b97043e57a22_video.html?utm_term=.05e9ff255689.

40. Matthew Feeney, “Border Patrol Seeking Facial Recognition Drones,” Cato at Liberty, April 10, 2017, https://www.cato.org/blog/border-patrol-seeking-facial-recognition-drones.

41. Government Accountability Office, “FBI Should Better En-sure Privacy and Accuracy,” GAO-16-267, May 2016, http://www.gao.gov/assets/680/677098.pdf.

42. Government Accountability Office, “FBI Should Better En-sure Privacy and Accuracy.”

43. Government Accountability Office, “FBI Should Better En-sure Privacy and Accuracy”; and Clare Garvie, Alvaro Bedoya, and Jonathan Frankle, “The Perpetual Line-Up,” Georgetown Law Center on Privacy and Technology, October 18, 2016, https://www.perpetuallineup.org.

44. Government Accountability Office, “FBI Should Better En-sure Privacy and Accuracy.”

45. Government Accountability Office, “FBI Should Better En-sure Privacy and Accuracy.”

46. Clare Garvie and Jonathan Frankle, “Facial-Recognition Software Might Have a Racial Bias Problem,” The Atlan-tic, April 7, 2016, https://www.theatlantic.com/technology/ archive/2016/04/the-underlying-bias-of-facial-recognition-systems/476991/; Steve Lohr, “Facial Recognition Is Accurate, if You’re a White Guy,” New York Times, February 9, 2018, https://www.nytimes.com/2018/02/09/technology/facial-recognition-race-artificial-intelligence.html; and Brendan Klare et al., “Face Recognition Performance: Role of Demographic Information,” IEEE Transactions on Information Forensics and Security 7, no. 6 (December 2012): 1789–1801, https://assets.documentcloud.org/documents/2850196/Face-Recognition-Performance-Role-of-Demographic.pdf.

47. Department of Homeland Security, “Privacy Impact Assess-ment for the Aircraft Systems.”

48. Craig Whitlock and Craig Timberg, “Border-Patrol Drones Being Borrowed by Other Agencies More Often than Pre-viously Known,” Washington Post, January 14, 2014, https://www.washingtonpost.com/world/national-security/border-patrol-drones-being-borrowed-by-other-agencies-more-often-than-previously-known/2014/01/14/5f987af0-7d49 -11e3-9556-4a4bf7bcbd84_story.html?utm_term=.113befbca5f4.

49. Government Accountability Office, “Additional Actions Needed to Strengthen.”

50. Government Accountability Office, “Additional Actions Needed to Strengthen,” p. 9.

51. “2 percent of Predator B flight hours from fiscal years 2013 through 2016 were attributed to support for state and local gov-ernment agencies.” Two percent of 20,780 hours is 415.6 hours. See Government Accountability Office, “Additional Actions Need ed to Strengthen,” p. 15.

52. Michael Peck, “Predator Drone Sends North Dakota Man to Jail,” Forbes, January 27, 2014, https://www.forbes.com/sites/michaelpeck/2014/01/27/predator-drone-sends-north-dakota-man-to-jail/#715b52bd324c; Brian Bennett, “Police Employ Pred-ator Drone Spy Planes on Home Front,” Los Angeles Times, De-cember 10, 2011, http://articles.latimes.com/2011/dec/10/nation/la-na-drone-arrest-20111211.

53. Jason Koebler, “North Dakota Man Sentenced to Jail in Con-troversial Drone-Arrest Case,” U.S. News and World Report, Janu-ary 15, 2014, https://www.usnews.com/news/articles/2014/01/15/north-dakota-man-sentenced-to-jail-in-controversial-drone- arrest-case.

54. Department of Homeland Security, “Privacy Impact Assess-ment for the Aircraft Systems,” DHS/CBP/PIA-018, Septem-ber 9, 2013, https://www.dhs.gov/sites/default/files/publications/privacy-pia-cbp-aircraft-systems-20130926.pdf; Department of Homeland Security, “Best Practices.”

55. Texas’s drone warrant requirement is subject to exceptions. For instance, Texas law enforcement can forgo a warrant if they are “in immediate pursuit of a person law enforcement officers have reasonable suspicion or probable cause to suspect has com-mitted an offense, not including misdemeanors or offenses pun-ishable by a fine only.” Texas Gov. Code Ann. § 423.002 (2013). In

9

2015, the New Mexico Supreme Court did rule that warrantless helicopter surveillance that caused disruption to the property be-ing surveilled did violate the Fourth Amendment. State v. Davis, 2015–S–1–SC–34548: “prolonged hovering close enough to the ground to cause interference with Davis’ property transformed this surveillance from a lawful observation of an area left open to public view to an unconstitutional intrusion into Davis’ expec-tation of privacy.” Allie Bohm, “Status of 2014 Domestic Drone

Legislation in the States,” ACLU, April 22, 2014, https://www.aclu.org/blog/status-2014-domestic-drone-legislation-states.

56. Didi Kirsten Tatlow, “China Said to Deploy Drones af-ter Unrest in Xinjiang,” New York Times, August 19, 2014, https://sinosphere.blogs.nytimes.com/2014/08/19/china-said-to-deploy-drones-after-unrest-in-xinjiang/?_php=true&_type=blogs&_r=0.

The views expressed in this paper are those of the author(s) and should not be attributed to the Cato Institute, its trustees, its Sponsors, or any other person or organization. Nothing in this paper should be construed as an attempt to aid or hinder the passage of any bill before Congress. Copyright © 2018 Cato Institute. This work by Cato Institute is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

RELATED PUBLICATIONS FROM THE CATO INSTITUTE

Extreme Vetting of Immigrants: Estimating Terrorism Vetting Failures by David Bier, Cato Institute Policy Analysis no. 838, April 17, 2018

Do Immigration Enforcement Programs Reduce Crime? Evidence from the 287(g) Program in North Carolina by Andrew Forrester and Alex Nowrasteh, Cato Institute Working Paper no. 52, April 11, 2018

How Mass Immigration Affects Countries with Weak Economic Institutions: A Natural Experiment in Jordan by Alex Nowrasteh, Andrew Forrester, and Cole Blondin, Cato Institute Working Paper no. 51, March 20, 2018

Criminal Immigrants in Texas: Illegal Immigrant Conviction and Arrest Rates for Homicide, Sexual Assault, Larceny, and Other Crimes by Alex Nowrasteh, Cato Institute Immigration Research and Policy Brief no. 4, February 26, 2018

Unwelcome Guests? The Effects of Refugees on the Educational Outcomes of Incumbent Students by David N. Figlio and Umut Özek, Research Briefs in Economic Policy no. 95, January 10, 2018

Border Patrol Termination Rates: Discipline and Performance Problems Signal Need for Reform by Alex Nowrasteh, Cato Institute Policy Analysis no. 825, November 2, 2017

The DREAMer Incarcerration Rate by Michelangelo Landgrave and Alex Nowrasteh, Cato Institute Immigration Research and Policy Brief no. 3, August 30, 2017

The Deep Roots of Economic Development in the U.S. States by Ryan Murphy and Alex Nowrasteh, Cato Institute Working Paper no. 44, May 19, 2017

State-Sponsored Visas: New Bill Lets States Invite Foreign Workers, Entrepreneurs, and Investors by David Bier, Cato Institute Immigration Research and Policy Brief no. 2, May 11, 2017

Criminal Immigrants: Their Numbers, Demographics, and Countries of Origin by Michelangelo Landgrave and Alex Nowrasteh, Cato Institute Immigration Research and Policy Brief no. 1, March 15, 2017

Immigrants Assimilate into the Political Mainstream by Alex Nowrasteh and Sam Wilson, Cato Institute Economic Development Bulletin no. 27, January 19, 2017

Surveillance Takes Wing: Privacy in the Age of Police Drones by Matthew Feeney, Cato Institute Policy Analysis no. 807, December 13, 2016

Terrorism and Immigration: A Risk Analysis by Alex Nowrasteh, Cato Institute Policy Analysis no. 798, September 13, 2016