The Perception of Biometric Technology: A Survey

The Perception of Biometric Technology: A Survey Stephen J. Elliott, Ph.D. Department of Industrial Technology, Purdue University West Lafayette, Indiana, USA elliottgpurdue.edu Sarah A. Massie Department of Industrial Technology, Purdue University West Lafayette, Indiana, USA smassiegpurdue.edu Mathias J. Sutton Department of Industrial Technology, Purdue University West Lafayette, Indiana, USA suttonmjgpurdue.edu Abstract- This paper outlines the perceptions of 391 individuals on issues relating to biometric technology. The results II. METHODOLOGY from this survey are broadly in line with other consumer perception surveys conducted in 2003 and 2005, despite the The surve was conducted over a four-month period increase in numbers of applications and deployments that affect in March 2006. The survey took into account other beginnyng individuals. Noticeably, 25% of the respondents had not heard of begininng inlmar 2 te survey i ontoter biometrics. Those who had heard of biometrics expressed several workscipnth fivan noted sun ie s by. posed concerns, including cleanliness of the devices, safety (with respect participantsin various research studies undertaken by members to iris and retinal identification), applications for which their of the Biometric Standards, Performance, and Assurance biometrics were going to be used, and who would have access to Laboratory (BSPA) at Purdue University. Some of the their biometric information. Results demonstrated overwhelming questions included in the survey resulted from interaction with support for biometrics applications involving law enforcement members of the general public who had visited the lab or and obtaining passports, while applications involving time and attended any of the many activities that the laboratory attendance tracking and access to public schools ranked lowest undertakes each year. Questions resulting from interaction with on the list. the general public focused on the safety of the technology, whether the iris camera uses lasers, and whether biometric Keywords- biometrics technologies are easily fooled. In many cases, the individuals appeared to be more interested in understanding the limitations and security implications of the technology than in I. INTRODUCTION understanding the specific benefits that the technology might The challenge of understanding public perception of provide. We wanted to know whether this greater interest in the biometrics usage is not new. In fact, many surveys have been technology than in its potential benefits was representative of conducted to assess the public's perception of biometric the public at large. The application questions were designed to deployments, as governments, industry, and the private sector address situations where biometrics are already being used, or see the benefits of using biometric technology. One study could be in used in the not-too-distant future. Forty-seven projected revenues to grow from $2.1B in 2006 to $5.7B in questions were generated covering topics that included 2010 [1], mainly due to incorporating biometrics in large-scale respondents' history of usage of biometric devices, their immigration programs such as the U.S. VISIT program (United opinions relative to their interaction with the devices; States); CANPASS (Canada); Ben Gurion Airport (Israel); and perceptions of the technology, and the perceived acceptability Schipol Airport (Amsterdam). Biometrics applications for log- of biometric technology in several different applications. in and access control are widely employed by the banking The survey was conducted both online and on paper. The sector, at ATM and advanced financial kiosks, for point-of-sale survey was distributed electronically to a number of sources in payment, and on personal consumer electronics (laptops, PCs, order to maximize the coverage of the potential audience. and tablets). Sources included listservs of groups and organizations involved Similar to previous surveys, the purpose of this study was in information assurance and security; students, faculty, and to ascertain individuals' perceptions of biometrics. In staff at a historically black institution; a Fulbright Scholar particular, we sought to understand whether differences in listserv; international student listserv; students at Kent State biometric technology's acceptability could be identified across University who were non-technology majors, individuals who genders, ethnicities, income levels, educational levels, and had visited the BSPA lab; individuals who enrolled in ages. biometric testing initiatives through advertisements in the local media, as well as random surveys conducted on Purdue University's main campus in West Lafayette, Indiana. 1-4244-1300-1/07/$25.OO 2007 IEEE 259

SURVEY PARTICIPANTS A total of 391 completed surveys were returned an d and seqe analys. that respondents 4900o were females. Table 1 shows that mostweres declared themselves to be white (770), followed by African American (70), Asian (7o), Native American (20), and Hispanic (40), and other (3 ). Approximately 480 of the survey respondents eared under $10,000 per year, 180 between $10,000 and $30,000; 11% between $30,000 and $50,000, and 8% earned over $90,000. A total of 51% had completed some college, 420 had completed college, and 80 had completed high school but not attended college. III. TABLE I. White 77 77 7 2 Hispanic 4 Some college Completed college 51 Other Completed high school $10,000-$30,000 $30,000-$50,000 42 48 18 11 m $70,000-$90,000 6 $50,000-$70,000 How physically confortable was the biometric device 10o 806040-9 20 $-1 1-4244-1300-1/07/$25.OO 2007 IEEE 1 2 3 How physicaly comfortable 4 5 Figure 2 depicts whether the individuals had any concerns about using the biometrics technologies for the particular application that they used the biometric device in the previous question. Of the individuals who reported having used biometric technologies, 27.50 of the respondents said that the use was mandatory, 43% was voluntary, and 29.5% responded that they did not know whether use of the technology was mandatory. 140D A co ncerns with biometrics Any 120D RESPONDENTS' HISTORY WITH BIOMETRIC MODALITIES Iris 0 F 1 R p In Figure 1, "-1" denotes not applicable, 1 is "very uncomfortable" through to "5" denoted as "very comfortable". Understanding Biometrics in General The introductory questions addressed whether the respondents had heard of biometrics prior to taking the survey. Respondents were provided a definition of biometric technologies because previous studies indicated that survey respondents often had problems in understanding the concepts without a definition ([3] and [4]). Out of the group of 391 respondents, 75.4% had read or heard about biometrics, while the remainder (24.6%) had not. Table 2 below shows the distribution of biometric modalities with which respondents had experience. Signature Voice Other number of Credit Union, hand geometry readers across the Purdue campus, hand geometry readers for access control and time and attendance, fingerprint recognition devices at local tanning salons, voice recognition systems in cell phone and signature collection components at point-ofapplications, sale payment devices in many local retail establishments. Respondents who had previously used biometric technologies were asked about their opinion on such technologies. Questions targeted whether respondents were physically comfortable in using the biometric device, whether they had any concerns about the use of the technology, and whether the use of the biometric technology was mandatory. Figure 1 depicts the respondents' reports of their physical comfort when using a biometric device. A. Fingerprint Hand Geometry a University biometrics in general; understanding of perceptions of biometrics with regard to vulnerability, privacy, and cleanliness; and suitability of use in particular applications, Biometric Modality used in locations for access control. These instances include fingerprint IV. SURVEY ORGANIZATION The survey was organized in three parts: understanding of TABLE 11. are recognition devices at ATMs in the Purdue Employee Federal 3 8 <$10,000 Income from the local area, where biometrics RESPONDENTS' DEMOGRAPHICS Whriten American 7merican African Ethnicity Asian Native American Education A total of 294 individuals had used various biometrics. This number is not surprising, in that many of the respondents were 800 60I Percentage 36 4020 11 Ncncern with biometrics 8 11 21 3 Figure 2. Individuals' concerns regarding use of biometrics 260

B. Perceptions regarding biometric technology Approximately 50% of the easy-to-fool group was concerned The general population seemed to have questions about or somewhat concerned with stealing body parts, and 86.7% biometric technology. These questions often relate to the were somewhat unconcerned or not concerned at all with security of such devices, how easy they are to break into or biometrics; this matter is discussed in greater detail in section E "spoof," and the privacy of such technology. The survey below. included six questions that the survey administrators had heard and asked repeatedly: D. Cleanliness * whether biometrics When can identify someone testing and using biometric devices, especially hand accurately, whether biometrics are easy to fool, geometry and fingerprint recognition systems, there were often comments on the cleanliness of the devices. The respondents of * whether fingerprint and hand geometry machines this survey did not provide either an overwhelmingly positive are sanitary, or negative response relating to the cleanliness of a biometric device. Approximately I11% commented that the devices were * whether iris or retial identification devices are very sanitary; however, 27.11% thought opposite, that the dangerous, devices were somewhat unsanitary. Approximately 75% of * whether biometrics would be used widely in the those surveyed thought it was somewhat sanitary, somewhat future, unsanitary, or neutral. * whether biometrics are important for homeland E. Safety security, The majority of respondents thought that iris and retinal * whether biometrics threaten or enhance privacy. identification devices were very safe, safe, or neutral. When asked whether they thought that, in the future, criminals would Individuals responded to these questions using a five-point steal body parts in order to circumvent a biometric device, 46% unidimensionatscale, which listed alternatives in five different of respondents were concerned. Of those concerned or very categories; points ranged from "not applicable" to "very concerned about the theft of body parts, 11% were age 40 and concerned." older, and 9.4% had only a high-school education. When comparing concern of body part theft versus ethnicity, annual C. Accuracy income, and age, Chi-Square analysis revealed only annual Approximately 87% of the individuals surveyed indicated income was significantly related (p-value = 0.045). The that biometric technology could identify individuals accurately. analysis did not include those who were very concerned of When asked whether biometrics were "easy to fool," 25.3% of body part theft due to a low expected frequency. respondents indicated that they were "very difficult" to fool, 25.3% noted that they were "difficult" to fool, and 21% were F. Frequency offuture use neutral on the subject. The remainder, approximately 5%, Approximately 95% of respondents noted that they thought thought that biometrics was easy to fool. biometric technologies would be widespread within the next Of those who felt biometrics were easy to fool (easy-to-fool decade, and 88.7% thought that biometrics were an important group), 87.5% had completed some or all college, 62.5% had tool for homeland security applications. When asked whether an annual income less than $30K, and 81.3% were males. The they thought that biometrics would enhance or threaten easy-to-fool group felt that biometrics were appropriate for the privacy, 26.3% were neutral, 34.8% were either concerned or following applications: identifying arrested persons (93.8%), very concerned, and 26% were either not concerned or hardly purchasing a gun (87.5%), obtaining a passport (81.3%), concerned. identifying people at border crossings (68.8%), preventing welfare fraud (68.8%), obtaining a national ID card (68.8%), G. Respondent Demographics safeguarding medical records (62.5%), securing a cell phone, In order to determine whether there was a significant PDA, or other mobile device (62.5%), obtaining a drivers relationship between demographic factors and respondents' license (62.5%), conducting ATM transactions (50%), and perceptions of biometrics, a Kruskal-Wallis analysis was randomly scanning crowds in search of wanted criminals at performed. Table 3 shows the results for a = 0.05. Bold values public events (50%). It is interesting to note, though the reason indicate significant relationships. is not clearly understood at this time, that the easy-to-fool group found biometrics appropriate for use in so many applications. Approximately 75%0 of the easy-to-fool group felt biometrics were somewhat or very accurate; 43 0 felt biometrics somewhat threatened personal privacy; and 68.8% believed biometrics were either a somewhat or very important tool for homeland security. Approximately 8100 felt biometrics would either probably or definitely be widely used in the future and 43.50/0 felt biometrics were somewhat or very unsanitary. 1-4244-1300-1/07/$25.OO 2007 IEEE 261

TABLE IV. BIOMETRIC APPLICATION SUITABILITY, AGGREGATED OPINIONS TABLE III. PERCEPTION QUESTIONS AND DEMOGRAPHIC VARIABLES, ApplicationPercentage KRUSKAL-WALLIS ANALYSIS Yes No p-value Identification of arrested people 92 7 Perception Gender Education Income Age Ethnicity Obtaining passports 91 8 Questions Can biometrics Purchasing a gun 84 15 identify 0.026 0.524 0.418 0.534 0.342 Obtaining a national identification card 82 17 accurately? Entering a government building 70 30 Are biometrics 0.921 0.992 0.778 0.498 0.061 Obtaining a drivers license easy to fool? Otiigadieslcne68 29 Are fingerprints Preventing welfare fraud 68 31 andghand 0.256 0.279 0.293 0.470 0.554 ID verification when using a credit card 67 32 sanitary? Safeguarding medical records 67 32 Is iris Checking in for a flight 65 35 recognition or retinal 0.31 0.182 0.044 0.127 0.164 Scanning public places 62 37 recognition dangerous? Making an ATM transaction 61 38 Wouldcriminals 0.139 0.088 0.017 0.565 0.020 Opening a bank account 58 42 steal body parts. Background check for employment 58 41 Would biometrics be 0.846 0.206 0.411 0.297 0.840 Voting in a national election 55 45 used widely? Logging into a computer at work 52 47 Are biometrics Payment authorization for online transaction 50 50 important for 0.659 0.375 0.817 0.138 0.222 homeland Securing a cell phone or PDA 47 53 security? Scanning for potential gamblers 36 64 Do biometrics threaten or 0.067 0.405 0.455 0.015 0.227 Entering a public school 32 67 enhance privacy? Logging into a computer at home 32 68 Time and attendance at work 30 70 Interestingly, education level was not significantly related Renting a vehicle 26 74 to any of the perception questions. Gender was significantly related to biometric accuracy (females thought it was more accurate than males). Income was significantly related to Respondents widely agreed that biometric technologies retinal device safety and whether criminals would steal body should be used to identify arrested individuals, obtain parts, although the two income groups most concerned with passports, purchase a gun, and obtain a national identification both these issues were the $70,000-$90,00 and <$10,000. Age card. Conversely, respondents disagreed about using biometrics was significantly related to privacy concerns. Out of all of the to rent a vehicle, maintain records about time and attendance at respondents, those older than 35 thought that biometrics would work, log on to a computer at home, and enter a public school. threaten privacy, whereas those younger than 35 believed that biometrics would enhance privacy. Finally, ethnicity was also A. Understanding Biometrics in General significantly related respondents' concerns about criminals...s stealing body parts; in fact all ethnicities except for whites had Whenieamionin temsuiability of biometicsnto v arious wiel vayn opnin on thsqesin applications by demographic, age was significantly related to obtaining a drivers license (p-value =0.017), entering a public school (p-value = 0.00 1), entering a government building (pvalue = 0.0 15), logging on to a computer at home (p-value = 0.027), logging on to a computer at work (p-value 0.027), Twenty-three questions were asked to solicit the maintaining time and attendance records (p-value 0.00 1), respondent's opinions on the suitability of using biometrics in renting a vehicle (p-value = 0.08), and authorizing payment certain applications. Table 4 shows a summary of the for online transactions (p-value =0.027). responses. 1-4244-1300-1/07/$25.OO 2007 IEEE 262

list as doing a good job of verifying identity. The same study found the least trusted organization to be police or law enforcement. Both [5] and [6] agree that the public is willing to use biometrics for entry and exit into countries. However in our study, the public feels that criminals and terrorists could still find a way to outsmart the system. On the other hand, research conducted on behalf of Citizenship and Immigration Canada found that Canadians feel the use of biometrics would make it more difficult for illegal immigrants to enter the country as well as increase difficulty for terrorists to operate within Canadian borders [3]. The same participants were willing to accept a different form of national identification that includes biometrics. Other concerns involve impjlementation cost and effectiveness of identification ([3] and[4i]) B. Significance ofrespondents 'Ethnicity Ethnicity was significantly related to how respondents felt about using biometrics to control entering a public school (p value = 0.026), performing background checks for employment (p-value = 0.001), and checking in for a flight (p-value 0.020). C. Significance ofrespondent's Gender Gender was significantly related to how respondents felt about logging on to a computer at home (p-value =0.010). D. Sign ificance ofrespondent's Education Respondents' concerns about utilizing biometrics to control logging on to a computer at work (p-value 0.040), securing a cell phone or PDA (p-value = 0.013), maintaining time and attendance records (p-value = 0.0001), and performing background checks for employment (p-value = 0.001) were significantly related to education. = There seems to be no standard in terms of the demographic information collected from respondents. Typically, only age and gender were consistently requested as part of questions asked of participants. Ponemon requested much more detail related to technological adaptivity, use of Internet, education ~~~level ' and income [6]. The author also included education levels, age ranges, and income level. Moody provided significance and correlation to both age and gender. In this study, gender played no role; however, older respondents were found to be less likely to use biometrics for online applications but more likely to use biometrics at ATMs [5]. E. Sign ificance ofrespondent 's Income..... to control Respondents' concerns about utilizing biometrics logging on to a computer at work (p-value =0.026) and securing a cell phone or PDA (p-value 0.024) were significantly related to income. VI. DISCUSSION OF CURRENT AND PREVIOUS SURVEY RESULTS The overall conclusion from the four studies is that there is a willingness to use biometrics for specific applications and that there is a definite lack of trust with particular organizations, mainly government affiliations, using biometrics. It appears that much work needs to be done in terms of educating users regarding the potential of biometrics. Misconceptions surrounding the technology must be overcome. In order to distinguish myth from fact within the public eye, academia and the biometrics community need to provide clear and concise information that the public can readily understand. Biometrics is becoming morepervasive within society. Ina survey of 1,003 people, almost three-fourths of those surveyed had heard or read about biometrics ([4]). However, a Canadian survey of 1,200 people showed that respondents had a very limited awareness of biometrics (under 10o) until they were prompted with a definition, at which point awareness rose to 5400O. In the Canadian survey, opinions of the respondents varied considerably, especially when their opinions were measured against age, education, and income. Citizen and Immigration Canada found that those with lower income and less education were more cynical about biometric technologies VII. REFERENCES [1] Anon., "Biometrics Market and Industry Report Addresses Biometric Investment Landscape Through 2010," (2006), retrieved [3]. For this survey, the respondents were informed of the definition of biometrics and provided with some examples of modalities (e.g., fingerprint recognition, hand geometry) at the beginning of the survey. In previous surveys, the most acceptable forms of biometrics were found to be the use of fingerprints ([2], [4], and [5]) and voice recognition ([6]). The modalities perceived as being most uncomfortable were iris and retinal biometrics ([5] and [6]). This is in line with our findings. While survey results have shown that many people feel January 5, 2007, http //www.biometricgroup.com/press releases/pr 2006 0.html - [2] Anon., "Biometrics and consumer applications," (n.d.), retrieved September 12, from 2006, http://www.findbiometrics.com/pages/consumer_articles/consumer html. I [3] Citizenship and Immigration Canada (2003), "Tracking public perceptions of biometrics," retrieved September 4, 2006, from http://www.cic.gc.ca/english/press/03/poll-biometric biometrics played a role in society and they will be widely used by the end of the decade ([3]), these surveys have also highlighted that people are concerned about privacy and the potential for misuse of information, especially by government agencies ([4] and [5]). According to [6], banking institutions are most trusted to manage identity credentials, perhaps because they are currently perceived as being first on a top-ten 1-4244-1300-1/07/$25.OO 2007 IEEE -r-2 from BMIRo201 -B [4] Anon., "Consumer attitudes about biometrics in ID documents," (August 2005), retrieved September 4, 2006, from ht:/w.rseogpfboercsuvpf 263

[5] J. Moody, "Public perceptions of biometric devices: The effect of misinformation on acceptance and use," in Proceedings of the Informing Science and Information Technology Education Joint Conference, Rockhampton, QLD Australia, June 25-28, 2004. [6] L. Ponemon, "Global study on the public's perceptions about identity management," submitted for publication by Unisys Corporation, April 26, 2006, retrieved on December 28, 2006 from http://www.unisys.com/products/news a events/all news/04268651i htm. 1-4244-1300-1/07/$25.OO 2007 IEEE 264