US 20050216953 A1
A system for verifying identities comprising an enrollment system having: (i) at least one alphanumeric input device; (ii) at least one biometric input device; (iii) at least one header file database having a plurality of identities; (iv) at least one search engine, said search engine in communication with said header file database such that said search engine first receives an alphanumeric data signal which has been input into said alphanumeric input device by the user, and then searches said database for identities that match the alphanumeric data according to a predetermined first set of criteria; (v) a processor to score the set of identities matched by said search engine according to a predetermined second set of criteria, said processor capable of determining the acceptability or unacceptability of said user's input alphanumeric data based on said score; and (vi) an identity escrow database which is in communication with said processor and receives from said unit an approved identity data signal based on the acceptability of the score, said escrow database additionally in communication with said biometric input device capable of receiving at least one biometric identity data signal input by the user to said biometric input device, said escrow database further comprising means for coupling the approved identity data signal and the biometric identity data signal to create at least one subfile within the escrow database for each user comprising the approved identity data signal and the biometric data signal. The present invention further comprises a verification system for verifying the identity of said user after the user has enrolled in the enrollment system. The verification system has means for processing a second input biometric data signal input by the user to the biometric input device to match the user's preexisting biometric data in said escrow database according to a predetermined third set of criteria. The verification system has an output device for transmitting to a third party whether or not a match was located within said escrow database for said user. The present invention further relates to a method and system for providing a warranty against the theft of a user's identity.
21. A system for analyzing identity records to detect stolen identities, comprising:
at least one alphanumeric input device;
at least one header file database having a plurality of identities;
a display device;
at least one search engine, said search engine in communication with said header file database such that said engine receives an alphanumeric data signal which has been input into said alphanumeric input device by the user, and then searches said database for identities that match the alphanumeric data according to a predetermined first set of criteria, and retrieves data for each matched identity including an identity usage type identifier and a time stamp; and
a processor to score the set of identities matched by said search engine according to a predetermined second set of criteria, said processor capable of determining the acceptability or unacceptability of said user's input alphanumeric data based on said score and preparing data for the graphical display of said user's input alphanumeric data and said matched identity data on the display device, such that each of said user's input alphanumeric data and said matched identity data is graphically portrayed according to usage type and according to a time sequence and acceptability score determined as a function of data associated with a first usage of said identity.
This invention relates to a system and method for establishing an identity of a user and, more particularly, to a system and process for identifying a user to electronic systems such as processors, computers, and computer and electronic networks.
Prior art methods of computer and internet security such as cryptographic processes, tokens, dongles, so-called “uncopyable media,” and various executable software protection schemes fail to prevent identity fraud. Such methods are incapable of ensuring that the person or entity at each end of a transaction is who he says he is. Because of the anonymous nature of the internet, the security of e-commerce-related information and transactions is a serious problem. At the center of the problem are those individuals who steal other persons' identities so as to perform fraud, pranks, vandalism, espionage and other illegitimate activities. Thus, the predominant internet security issue is identity authentication.
While authentication takes various forms, authentication of individuals is particularly desirable. Authentication is directed to verifying that the individual seeking access to and/or through a server is in fact who that individual claims to be, and not an impersonator. This authentication relies on verification being performed at or above a predetermined minimum level of confidence. At the same time, authentication is generally an early hurdle that the individual must clear to conduct internet transactions with the server.
The traditional method for authenticating individuals has relied on secret passwords. Password-only authentication can be implemented entirely in software. However, password-only authentication has a number of disadvantages. For example, a password's viability is enhanced, among other ways, by increasing its length, by controlling its composition and by it being frequently changed. This, however, is cumbersome and, additionally, passwords can be lost or stolen, particularly written passwords. Passwords can be inadvertently disclosed to crackers via various ploys, such as observing the password's entry on a keyboard. Moreover, passwords can be intercepted as they are transported from the user to the desired server. Consequently, password-only authentication fails to provide adequate security.
Internet-based applications are flooding into areas that can benefit from enhanced security. Examples of such web-based applications include commercial transactions (e.g., loans and the purchase and sale of goods), banking transactions (e.g., electronic finds transfer), and medical transactions (e.g., provision of medical records in emergency situations). The internet is redefining commerce by eliminating the constraints of time and distance. World internet commerce sales are projected to reach between $1.7 and $3.5 trillion by the year 2003 (Source: Forrester Research, Inc.). Identity information and the authentication thereof will drive this explosion. However, many are uncomfortable with the current privacy protections. Although the Merchant's Association reports that e-business is growing by 200% annually, only about 5% of consumers visiting a website actually make purchases. The primary reason for this discrepancy is consumers' concern about privacy and online security. Has Business Accepted the Self-Regulation Challenge? Federal Trade Commissioner Mozelle Thompson, Privacy in American Business, Fifth Annual Conference Journal a (February/March 1999). Moreover, a recent Business Week/Harris Poll confirms that almost two-thirds of non-internet users would be more likely to start using the internet if the privacy of their personal information and communications would be protected, and that privacy was the primary reason individuals are choosing to stay off the internet, coming in well ahead of cost, concerns with complicated technology, and unsolicited commercial e-mail. Business Week/Hams Poll on Online Insecurity, Lewis Harris & Associates, Inc., New York, March 1998.
Additionally, a 1996 Harris Poll reported that 24% of Americans have personally experienced a privacy invasion, which is up from 19% in 1978. The same survey found that 80% of Americans felt that consumers have lost all control over how personal information about them is circulated and used by third parties. Equifax/Harris Consumer Privacy Survey, Lewis Harris & Associates Inc., New York, February 1996. Indeed, such fears have been confirmed by actual incidences of identity theft reported by the media. See, e.g., Hacker Discloses 350,000 Numbers: Web Retailer's Credit Security Breached, Chicago Tribune, Business, p. 1 (Jan. 11, 2000); Doubts Tiggered Over Web Shopping, Assoc. Press, Jan. 20, 2000 (A “19-year-old Russian” claimed to have stolen 300,000 credit card numbers by exploiting a flaw in CD Universe's System). Accordingly, there is an acute need in the art for a system and method for verifying identity which goes beyond known systems and methods where a user's submitted identity information is not cross-checked against a database of identity information to halt fraud and/or determine the likelihood of an attempt to use fraudulent information.
Obviously, there is a multitude of instances where it is necessary to verify that an individual requesting access to a service, an e-commerce transaction, or a facility is in fact authorized to access the service, execute the transaction or enter the facility. For example, such services include banking services, or telephone services, while the facilities may be for example banks, laboratories, computer systems, or database systems. In such situations, users typically have to write down, present a card, type or key in certain information in order to send an order, make a request, obtain a service, perform a transaction, transmit a message, or enter a facility. Verification or authentication of a customer prior to obtaining access to such services or facilities typically relies essentially on the customer's knowledge of passwords or personal identification numbers (PINs), possession of a card or token, or by the customer interfacing with a remote operator who verifies the customer's knowledge of information such as name, address, Social Security number, city or date of birth, mother's maiden name, etc. In some special transactions, handwriting recognition or signature verification is also employed.
However, such conventional techniques present many drawbacks. First, information typically used to verify a user's identity may be lost or stolen and, with existing technology, a criminal may find it easy to obtain such personal information such as the Social Security number, mother's maiden name or date of birth of his intended target. The shortcomings inherent with the conventional security measures have prompted an increasing interest in biometric security technology, i.e., verifying a person's identity by personal biological characteristics, such as voice printing, finger printing, iris scans, etc. However, even with biometric systems of the prior art, no attempt is made to cross-reference the user's alphanumeric identity data (i.e., name, address, Social Security number, etc.) against a database of identities which can determine, to a high degree of certainty, whether the alphanumeric identity data being offered with the biometric identity data is suspicious and/or subject to fraud. Without such cross-checking, a criminal submitting a biometric exemplar together with stolen alphanumeric identity data cannot be recognized as the fraud that he is by the anonymous computer systems which are so prevalent today.
Accordingly, a need exists for improved network and internet-based systems and methods to verify identities.
The system of the present invention provides an identity-based access control. It defines the authority and delegation of authority to access information, and provides for accountability. The present invention permits the ordering of goods and services in a secure manner over an open and anonymous environment such as the internet or other insecure network. The present invention further permits the payment for goods and services to be transmitted across an open network without fear of diversion to an unauthorized payee. It permits the delivery of intangible personal property and various electronic products in a secure fashion over such open networks. Additionally, the present invention permits the negotiation and formation of contracts in a secure manner over open networks; permits the conduct of auctions in a practical, reliable and trustworthy manner; permits the execution of guarantees in a trustworthy and reliable manner; permits the handling of various securities transactions, including stock purchases, in a secure fashion; and has the advantage of providing a consistent application programming interface which can be utilized in all types of transactions for ensuring security and authenticity of identities.
The invention is also directed to a method of conducting electronic commerce over an unsecured network by enrolling users in an infrastructure system to create an identity escrow or “virtual safety deposit box” for each user, and by verifying the authenticity of each electronic transaction by the user supplying a biometric sample with each transaction such that the biometric sample serves as the “key” to the user's identity escrow, and which causes the escrow to transmit a message to a third, party that the user is who he says he is.
These and other advantages and objects of the invention are achieved by providing a system for verifying identities comprising an enrollment system having: (i) at least one alphanumeric input device; (ii) at least one biometric input device; (iii) at least one header file database having a plurality of identities; (iv) at least one search engine, said search engine in communication with said header file database such that said search engine receives an alphanumeric data signal which has been input into said alphanumeric input device by the user, and then searches said database for identities that match the alphanumeric data according to a predetermined first set of criteria; (v) a processor to score the set of identities matched by said search engine according to a predetermined second set of criteria, said processor-capable of determining the acceptability or unacceptability of said user's input alphanumeric data based on said score; and (vi) an identity escrow database which is in communication with said processor and receives from said unit an approved identity data signal based on the acceptability of the score, said escrow database additionally in communication with said biometric input device capable of receiving at least one biometric identity data signal input by the user to said biometric input device, said escrow database further comprising means for coupling the approved identity data signal and the biometric identity data signal to create at least one subfile within the escrow database for each user comprising the approved identity data signal and the biometric data signal.
The present invention further comprises a verification system for verifying the identity of said user after the user has enrolled in the enrollment system. The verification system has means for processing a second biometric data signal input by the user to the biometric input device to determine a match of the user's preexisting biometric data in said escrow database according to a predetermined third set of criteria. Finally, the verification system has an output device for transmitting to a third party whether or not a match was located within said escrow database for said user.
The present invention additionally comprises a system and method for providing a warranty to users against identity theft.
Still other objects and advantages of the present invention will become readily apparent to those skilled in the art from the following drawings and detailed description, wherein only the preferred embodiment of the invention is shown and described simply by way of illustration of the best mode contemplated of carrying out the invention. As will be realized, the invention is capable of other and different embodiments, and its several details are capable of modifications in various obvious respects, all without departing from the invention. Accordingly, the drawings and description are to be regarded as illustrative in nature, and not as restrictive.
The principles and operation of the system and method of the present invention may be better understood with reference to the drawings and accompanying description. Fraud is epidemic and growing at five to ten times the rate of overall economic growth. This dramatic increase in growth can be largely attributed to the concomitant proliferation of digital systems and the naďve use of “fraud detection” systems.
Scam artists are knowledgeable about the detection systems and use the detection system's own logic to defeat them. These systems took the seemingly easy route in design and are built upon the detection of “suspect” transactions. Knowing which transactions might be “suspect” is only a starting point—not an end in itself. The best measure of such systems (frequently expensive neural network systems) is their early success followed by massive increases in fraud. This is much like the growth of a bacteria that becomes resistant to antibiotics. The present invention differs dramatically in its approach by concentrating on the perpetrator of the fraud rather than the fraudulent transaction. This results in increases in efficiency and effectiveness at far lower cost.
1. System Components.
A. Enrollment System Components.
The system of
Portable storage medium drive 22 operates in conjunction with a portable non-volatile storage medium, such as a floppy disk, or other computer-readable medium, to input and output data and code to and from the computer system of
Input device(s) 20 provide a portion of a user interface. Input device(s) 20 may include an alphanumeric keypad 46 (
The devices contained in the computer system of
Alternative embodiments of the use of the method of the present invention in conjunction with the computer system further include using other display means for the monitor, such as CRT display, LCD display, projection displays, or the like. Likewise, any similar type of memory, other than memory 14, may be used. Other interface means, in addition to the component interfaces, may also be used including alphanumeric keypads, other key information or any pointing devices such as a mouse, trackball, stylus, and cursor or direction key.
In a further embodiment, the present invention also includes a computer program product which is a storage medium (media) having instructions stored thereon/in which can be used to program a computer to perform the method of interfacing of the present invention. The storage medium can include, but is not limited to, any type of disk including floppy disks, optical disks, DVD, CD ROMs, magnetic optical disks, RAMs, EPROM, EEPROM, magnetic or optical cards, or any type of media suitable for storing electronic instructions.
Stored on any one of the computer readable medium (media), the present invention includes software for controlling both the hardware of the general purpose/specialized computer or microprocessor, and for enabling the computer or microprocessor to interact with a human user or other mechanism utilizing the results of the present invention. Such software may include, but is not limited to, device drivers, operating systems and user applications. Ultimately, such computer readable media further includes software for performing the method of interfacing of the present invention as described above.
As illustrated in
Mass storage device or server 16, output device or server 32 and memory 14 may be implemented by one digital processor 12. In that case, consolidation, scheduling, initial and subsequent segmenting of customers and execution of working programs are accomplished through the one processor running the inventive software. Neural networks may be employed to operate on mass storage device 16 and portable storage medium drives 22 to learn each individual customer's purchasing behavior and segment customers and potential customers accordingly. As used herein, “customer(s)” is the person who is the subject of the identity verification.
Further, instead of the neural networks, a multiple regression correlation based on the customer data in devices 16 and 22 may be used for segmenting customers. Alternatively, rule based expert systems may be similarly employed in the present invention to provide dynamic behavioral segmentation.
The expert system comprises a set of decision rules which operate to “customize” the processing and output of the software system used in the enrollment and verification of each customer, based on certain customer-specific input data. The decision rules define the logic used to make decisions which, in turn, becomes additional inputs to the software system for the purpose of customizing the output presented to the customer.
In a further example, and as more fully described below, the invention may be employed on the internet 40. A website can be created with a home page with topic selections and links (e.g., Hypertext HTML technology) to appropriate software programs 42 triggering questionnaires, identity verification and the transmission of information. In particular, for each topic selection there is a respective hyperlink to a software program 42 and optionally an event 44 for initiating the program 42. Upon the customer's selection of a topic from the home page, the present invention applies the linked event 44, if any. If the criteria of the event 44 are met (or if there is no initiating event 44), then the present invention executes the corresponding working program 42. This results in the working program 42 contents being transmitted online to the customer.
In another embodiment, the present inventive system 10 is based at a main site of a company. Different branches, departments and/or sites of the company may utilize the system from separate workstations as though they are the separate systems 10.
As further explained below, when search engine 54 builds keys for each user's identity data signal, different keys and weighting instructions are used for common names versus normal names. The primary difference is that the numeric component (i.e., SSN) of a common name record is weighed more heavily than the alphanumeric component as a predetermined first set of criteria. Consequently, the user's identity data signal is modified to identify it as an NN or CN to search engine 62.
Search engine 62 is comprised of search software known in the art, such as SSA-3 available from Search Software America. Search engine 62 is installed and configured to run on and sort header file database 60. Database 60 contains one or more header files available from the commercial credit bureaus such as, for example, Equifax or Deluxe Corporation. The header file contains identity information (i.e., names and Social Security numbers) of hundreds of millions of identity records. Because there are only about 160 to 180 million people in the United States who could be in database 60, most individuals appear in the database more than once.
Upon the initial installation of header files to database 60, search engine 62 first searches the header files against themselves to sort the similar or related identities into sets, and to locate sets of identities that contain suspicious variations of identity information. The matching identities of database 60 are scored (as explained below) according to a predetermined second set of criteria to determine the suspicious records.
B. System to Recognize Deceptive Identities.
After search engine 62 performs the initial sorting and scoring of all identity records, it is ready to be presented with a user's enrollment identity data record from filter 54. The data signal of the user's alphanumeric identity record from filter 54 is assigned its own set of keys (i.e., FN, MI, LN, SSN, NN, CN) in search engine 62's software. These keys are a predetermined set of criteria, which are used to search database 60 for matching identity records. The user's keys are compared with the keys in the credit bureau header file database 60, and the returned matching identities are scored as shown below to determine the deviation from the captured alphanumeric identity data.
As the scores increase, generally changes become subtler. As scores increase from 50 to 100 they represent changes in groups from two or more individuals to a single individual making suspicious changes. At 96 to 97 there is a transition from suspicious to innocent changes.
The following charts represent the relative distribution used to score the various data deviations and to determine the likelihood of a fraudulent manipulation.
C. Creation of Scoring Method.
To achieve the above scoring system, applicant employed several unique search strategies. The first strategy relied upon the importance of context in giving meaning to the content of the information. For example, if the content of the information is “9 to 5,” depending upon the contextual setting, the information may be one's working hours; or it might be the final score in the seventh game of the World Series. Context can dictate the meaning of the content. Thus, the context of the records derived from attempts to open checking accounts at banks was an important factor in how applicant analyzed the data.
The second strategy relates to the premise that information may be false and should be treated that way. Since applicant was looking for attempts to defraud banks, he assumed the professional criminal was likely lying to the bank about who the criminal was. This is fundamental because if the criminal were to identify himself as someone who defrauded banks, the bank would be not deal with him. The importance of this assumption of falsity for the information system was crucial. To prove that a criminal lied about his identity, applicant's goal was to find two identities purportedly belonging to the same person that were different, suspicious and mutually exclusive.
Conventional wisdom dictates that the more information fields analyzed, the better the result However, many of the information fields, such as addresses, drivers license numbers, phone numbers, etc., not only legitimately change over time, but are frequently obtained from unreliable sources. Dealing with unreliable and potentially innocent changes to information creates a burden and could well bias the final output of the system. Applicant focused entirely upon information that does not legitimately change over an individual's lifetime, except for very specific and known reasons, e.g., a woman changing her name when she marries. Alterations in the name fields and the Social Security number field could be assumed to be either innocent (e.g., the use of a nickname, or the adoption of a married name, or a typographical error), or suspicious. Therefore, applicant needed only to identify and screen out the known types of innocent and legitimate identity changes to be left with the suspicious changes.
Because there is a nearly infinite range of identity changes possible between one person manipulating their identity and the real identity of a third person, applicant needed to carefully calibrate his system to draw the line between intentional suspicious alterations and innocent third party identities (i.e., false positives). This iterative calibration resulted in the above-stated scoring grids.
The above scoring or benchmarking process is used to determine to a high degree of certainty whether a change in identity data is intended to deceive the recipient of the data. Applicant has determined that scores of 76 through 97 represent changes associated with an intent to deceive. Such determination about this scoring range resulted from applicant's analysis and measurement of fraudulent conduct in the United States. Using 1993 data from the Justice Department (Uniform Crime Report and affiliated documents) and census data, applicant constructed a model of fraud activity in the United States for 1993. Applicant plotted the number of fraud arrests per capita by state as illustrated in
To ensure that the phenomenon depicted in
In any event, analyzing 30 to 40 factors proved to create too much noise in the system and skewed the data Based on applicant's years of experience of being involved in fraud cases, he has observed that perpetrators lie, but lie very little. Perpetrators desire to tell as much truth as they possibly can to establish their credibility. Credibility is crucial to their success. Today, a substantial amount of fraud takes place in banks. To verify identities, banks go to large data managers such as Equifax, Trans Union, Experian, etc. Thus, the perpetrator must successfully weave his way through such systems. He can only do so by manipulating his identity. The perpetrator prefers to reuse the transposed numbers and letters of an altered identity in a way that they appear innocent and can be explained away (e.g., typographical errors). In order for the criminal to defeat the systems in place at the credit reporting agencies and other record providers, the criminal need only make a subtle change in identity. Each single identity has tens of thousands of variations that do not require a significant name change. Criminals are expert in digital systems (most fraud is perpetrated by organized criminal elements), and can easily defeat all of the screening systems of the prior art. Applicant targeted this specific behavior as a means for identifying suspicious conduct that is likely to lead to a fraud loss if unchecked.
In order to solve the problem of excess noise and false positives, applicant employed an iterative searching process to determine which changes are innocent and which ones are fraudulent. A database representing about 56 million records of checking account openings in roughly the same period (1993) as the fraud data was supplied by a subsidiary of Deluxe Corporation. Because the geographic market penetration of the Deluxe data did not perfectly match the actual census data, the fraud model was normalized using standard techniques to account for the difference. A name matching fuzzy logic software package was acquired from Search Software American (Name 3). This name matching software allows for the fuzzy association of both alpha and numeric information with a wide range of nearly infinitely tunable association algorithms tied to the various identity data in the system. The data was first grouped into sets of matching or related identities.
It was applicant's hypothesis that a curve of changes in the identity data (controlled for geographic location) would substantially match the curve of per capita fraud ranked by geographic location. Beginning with the default settings supplied by the search software, a curve of per capita associations (i.e., groupings of similar identities that represent the same digital identity) was generated and compared to the curve from the standardized (normalized) fraud model curve of
The above scoring charts and system resulted from that precise algorithm. To test applicant's results of a 94.5% correlation between the two curves, applicant compared the records of 188,602 individuals identified by applicant as having manipulated their identities in opening checking accounts with ChexSystems' comprehensive Reported Names Database and obtained 188,415 record matches. In addition, applicant obtained the identities used by known perpetrators of bank fraud from a bank customer. From this list, 21 identity records that matched the timeframe of applicant's test data were blindly distributed within the 56 million records used for system development. Applicant's random chances of identifying the test records were 21:56,000,000. Nevertheless, applicant successfully identified all 21 known perpetrators and identified no false positives. Applicant has further tested the system with successful results.
After a user's alphanumeric data is scored by search engine 62, the system of the present invention continues the inventive enrollment process. If the score is within the acceptable limits of confidence and the submitted identity data are not suspicious, then the record is transmitted via output file 64 and affirmative step 66 to approved file 70. The suspicious sets of identity data determined by the above process are output via output file 64 and transferred to derogatory file 72 in step 68. File 72 comprises alphanumeric subfile 74 and biometric subfile 76, and will typically contain approximately 10 million such sets of suspicious records and approximately 32 million identity records. Based on applicant's empirical data, these sets of identities have been intentionally manipulated by individuals for the purpose of deception. A copy of files 72, 74 and 76 are transmitted to database 60 as a folder to maintain the comprehensiveness of database 60 and to, ensure the most accurate scoring when a user enrolls. The non-suspicious sets of identities remain in database 60 and are copies of each other and innocent by nature.
The alphanumeric component of enrollment system 50 is illustrated by the following examples:
Enrolling for the first time is:
William Clinton's alphanumeric data reveal suspicious records according to applicant's scoring system. William Clinton is not approved and his record is transmitted to derogatory file 72 with a copy in a derogatory file folder in database 60. Thus, if William Clinton attempts to enroll again, the derogatory data located in database 60 will be triggered, and enrollment is denied.
The enrollment process continues with obtaining one or more biometric exemplars from the user using biometric input device 48. The system of the present invention can utilize any commercially available biometric, including face printing, voice printing, finger printing, retinal printing, iris printing and DNA mapping. Numerous biometric devices are known in the art, such as those available from Mytec Technologies of Toronto, Canada; Intelitrak Technologies of Austin, Tex., and National Registry, Inc. of Tampa, Fla. Additional information regarding biometrics can be found at the International Biometric Industry Association's website at www.ibia.org.
Most users will opt for the use of voice printing biometrics because no additional computer equipment is necessary if the user's computer includes a microphone, as is the case with most computers. Alternatively, a signal can be sent from system 50 to the user's computer to put the user's modem on hold, to permit the user to employ a telephone handset to deliver a voice sample in the enrollment and verification processes.
The biometric exemplar, in digital form as known in the art, is captured and transmitted via a communication network, preferably the internet, to the Central Biometrics Authority (CBA) 78, a system available from the International Biometrics Group (IBG), or other like system, to ensure that the particular biometric employed by a user meets the appropriate standard of sensitivity set for the type of device used and the type of transaction.
CBA 78 has the ability to process the information from all commercially available biometric technologies and devices in a manner that captures the setting of the devices and the degree of certainty that the device is capturing the biometric. This includes such things as the number of attempts to certify the biometric, whether the certification is contemporaneous or stored, etc. This information is then integrated to match the desired or required certainty level for the party requesting verification. For instance, an online stock brokerage may desire a very high level of certainty for a $100,000 stock trade, but an online bookstore might only require a 51% certainty level for a $10 sale.
Once the CBA 78 approves the submitted biometric, the biometric extract 80 is transmitted to and stored in biometric suspense buffer 82. If the CBA 78 rejects the biometric exemplar from device 48 in step 84, a signal is sent to the user from CBA 78 to resubmit another biometric exemplar. This process can repeat itself an unlimited number of times, although generally alter a user's biometric is rejected three times, CBA 78 will send the user a message informing her/him to employ a different device 48.
Bach biometric input device 48 is preferably linked to the alphanumeric input device 46 such that the biometric exemplar is identified as that of the user by name, address, Social Security number, password, etc. Thus, each biometric record in suspense buffer 86 has some alphanumeric data identifier in order for it to be matched with the user's alphanumeric identity record which was processed as explained above and either approved or rejected in steps 66 and 68, respectively.
After the biometric record is stored in buffer 86, buffer 86 then queries approved file 70 and derogatory file 72 for matching alphanumeric identity information. If a match is found in file 70, the approved alphanumeric identity record and the approved biometric record are coupled in storage device 88 and transmitted to identity escrow server or database 90 (“Virtual Safety Deposit Box”). Accordingly, each enrolled user's approved alphanumeric and biometric records are stored in identity escrow database 90 ready for use in the verification process. Bach user's records can be further coupled to ancillary database 92, which may contain any other personal information such as financial and medical records. Conversely, if a match is found in file 72, the user's biometric is coupled with the user's alphanumeric data from derogatory subfile 74 and transmitted and stored in biometric subfile 76.
D. Verification System.
2. System Operation.
A standard personal computer capable of internet access or a standard telephone or other known telecommunication device are compatible with the system of the present invention. Computer system 10, for example, can be employed. Computer system 16 further comprises biometric input device 48, which is a combination of hardware and software and enables the user to provide the biometric exemplar for enrollment into the system, but also the biometric samples for each use of verification system 100. Biometric input device 48 is selected from a wide variety of available technology, including fingerprint recordation devices, voice print in recorders, retinal image recorders, hand geometry recorders, and the like. Device 48 can be combined with a camera so that a photograph or digital image becomes a part of the biometric data record. The biometric data record is preferably encrypted to enhance security. Depending upon their construction and sensitivity, devices 48 vary in ability to resist fraud. Consequently, e-retailers may dictate more stringent device ID, parameters for high-valve, high-security transactions (e.g., stock trades), or less stringent parameters for low-value, low-security transaction (e.g., sale of $20.00 item). Because almost all biometric devices 48 have different sensitivities and error rates they thus have different software attendant thereto. As explained above, the system of the present invention communicates with the CBA 78 during enrollment and verification (as part of a predetermined third set of criteria) to ensure accuracy of the biometric sample.
To enroll in system 50, users provide alphanumeric identity data via the internet, by telephone, by mail, etc. System 50 (as well as system 100) assigns and encrypts this data, prepares forms for mailing and signature and communicates with third party networks. As described above in connection with enrollment system 50, such identity data is searched against existing data in database 60 and scored to determine the existence of suspicious identity records. All suspicious records are copied to derogatory file 72, and the users providing such information are declined from enrollment system 50 or are requested to provide additional information. Notice to such users is by e-mail or other known means. Consequently, enrollment includes a confirmation step, to a high degree of certainty, that the particular individual is who he/she says he/she is.
Enrollment further comprises acquiring from the user data representing one or more of the individual's biometric characteristics. After successful processing by CBA 78, such data is stored in biometric suspense buffer 86. Multiple biometric data samples are preferred so as to account for variations among the samples and to increase the reliability of authentication. As detailed above, in order for the system to be ready to verify a user's identity, the biometric data record is then stored in database 90 in association with the user's approved alphanumeric identity record of file 70.
Preferably, but not required, system 50 assigns each user an anonymous or shared personal identification number or name (“PIN”). This PIN is coupled to the user's alphanumeric and biometric records and is shared with many other users. The PIN may be coupled to such records by the party seeking verification, i.e., the merchant, financial institution or government. The purpose of using a shared PIN is that it speeds the verification process. For example, system 50 has assigned (or the user has chosen) “Sammy Sosa” as the user's PIN. If there are 9,999 other users utilizing “Sammy Sosa” as their PIN, system 100 only has to search 10,000 files during verification. The PIN may be in alphanumeric or digital voice form. Although the system of the present invention will function using an individualized, non-shared PIN, there are disadvantages besides slowing verification system 100 with which to contend. For example, the user may be in a location where her/his PIN can be seen overheard by third parties. Conversely, with a shared PIN, there is no need to keep it confidential.
Users can be individuals or business entities and, thus, enrollment can take many forms. For example, authorization can be restricted to one or more individuals within a particular organization.
During verification, verification system 100 receives encrypted alphanumeric and/or biometric identity data via devices 46 and/or 48. Such identity data can be transmitted to system 100 from virtually anywhere using known computer and telecommunications technology. Applications include e-commerce transactions where a website requests verification of the identity of one of its users prior to completing a transaction; home security systems, military security systems and countless other applications. Such third parties and e-commerce providers are preferably registered with system 100. Upon transmission, these data are decrypted by database 90 using a channel supporting secure socket layer (SSL) or some other security protocol. Database 90 then determines whether it has enrolled the claimed identity of the user seeking authentication. CBA 78 and database 90 preferably filter out unacceptable messages. Such messages can include those claiming an identity that does not agree with any records available in database 90, such as messages containing PINs or identity data of persons not enrolled in system 100, or messages from websites not registered with system 100.
In configurations using PINs, preferably shared PINs, database 90 preferably determines whether the transmitted PIN matches any of the PINs in database 90. Within the identities stored under the shared PIN, each biometric technology has varying capabilities of performing limited 1 to N searches.
The identity escrow server/database 90 compares the alphanumeric and biometric data (or the biometric data alone) of each transmitted message with the records stored in the database 90. Recall that the biometric transmission is first processed by CBA 78, although this step is not essential to successful operation. To verify the individual's claimed identity, database 90 typically compares the transmitted biometric data to the records of the enrolled individual.
A user's live data may fail to match exactly the user's biometric records. This occurs because acquisition of biometric data is subject to variations, both in the enrollment mode and in verification mode. Accordingly, CBA 78 and database 90 preferably employ comparison algorithms that do not require exact matches between the live data and records, but provide a high level of accuracy. The comparison algorithms generally are implemented to determine the statistical closeness of the live data to the records. In that implementation, database 90 produces an affirmative response in the event of a sufficient match, i.e., if the closeness determination yields a result that falls into a selected range of confidence determined to be acceptable. In contrast, database 90 produces a negative response if the result falls into a selected range of confidence determined to be unacceptable. Database 90 preferably also supports an indeterminate response (e.g., try again/provide more data) if the above ranges are not complementary and the result falls between such ranges. The indeterminate response can also result for other reasons, including that database 90 is down, busy or otherwise. The ranges of sensitivity used in the comparison algorithms may vary among biometric characteristics. For example, a range having high sensitivity may be selected for highly secure transactions and vice versa. In any case, optimum sensitivity selections are generally made to strike a balance between false positives and false negatives.
Following each verification, database 90 produces a response in steps 110, 112 and 114. Database 90 transmits the response to output device 116, the response being that the user's identity is or is not verified from processing steps 110 and 112, or that system 100 contains no data for such user from processing step 114. Output device 116, and system 100 for that matter, only transmits to third party 118 whether or not the user is verified or that no records exist. No identity data of any kind is ever ID transmitted to third party 118. Thus, verification system 100 is highly secure against hackers.
If output device 116 transmits a “No” message resulting from step 112, third party 118 and system 100 may employ a predetermined set of criteria to request the user to re-attempt verification. The number of re-attempts allowed may depend on the level of security for the particular transaction.
If, in step 114, no record is found in database 90, this message is transmitted by output device 116 to third party 118. Third party 118 may then electronically or telephonically direct the attempted user to enrollment system 50 through a different website or server, or enrollment system 50 may be integrated with third party 118's system so as to seamlessly enroll the user into system 50. Alternatively, third party 118 may provide the attempted user with a website address or link, telephone number or mailing address for direct contact with system 50.
Plural verification systems 100 and enrollment systems 50 can be employed such that verification can be specific for a particular third party 118. For example, a bank may operate a captive database 90 only for its customers who have registered with system 50 when opening their respective bank accounts. Conversely, in an open structure, database 90 can process verification requests from any source.
The identity verification system of the present invention can be employed for any type of transaction or event where verification of the user is necessary or helpful to prevent identity fraud. Such transactions may include any e-commerce transaction such as the purchase of goods using a credit card where, as part of the computerized credit card authorization process, the user is queried as to whether he/she has previously enrolled in system 50 and, if so, is asked to supply the necessary biometric to be processed through verification system 100. Once system 100 responds with a “Yes” to the credit card vendor, the vendor can authorize the charge and the retailer may complete the transaction, the vendor having received verification with a high level of confidence that the user of the credit card is the actual owner of the card.
The present invention further relates to a system a method for providing users of systems 50 and 100 with a warranty against the theft of their identity data. Warranty system 200 is linked to systems 50 and 100 by hyperlink or other telecommunication means. System 200 is similar to system 10 and comprises like components: Processor 202, mass storage device 204, memory 206, peripheral device 208, input 210, portable storage medium drives 212, graphics subsystem 214, audio subsystems 216, display 218, output device 220 connected by bus 222. System 200 may be further connected to a network 224, such as the internet 226 and third parties 228. As with system 10, events 230 trigger software 232 to activate and run system 200. Upon the request of a customer using systems 50 and 100, warranty coverage can be purchased. Such coverage protects the customer from the theft or misuse of his/her identity data used in systems 50 and 100.
3. User Interface.
As illustrated in
While the present invention has been described in connection with the preferred embodiments, it will be understood that modifications thereof within the above principles will be evident to those skilled in the art and, thus, the invention is not limited to the preferred embodiments but is intended to encompass such modifications and all equivalents thereto.