10.07.2015 Views

Report - Guardian

Report - Guardian

Report - Guardian

SHOW MORE
SHOW LESS
  • No tags were found...

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

The LSE Identity Project <strong>Report</strong>: June 2005 177According to one expert, our understanding of fingerprints “is dangerously flawed andrisks causing miscarriages of justice”. 434 Amongst the numerous cases of mistakenidentification through fingerprinting, that of Brandon Mayfield is indicative of the manyproblems in assessment and interpretation of fingerprint data.Following the Madrid Bombings of March 11 th 2004, Spanish National Police managedto lift a fingerprint from an unexploded bomb. Three highly skilled FBI fingerprintexperts declared that Oregon lawyer Brandon Mayfield's fingerprint was a match to thecrime scene sample. U.S. officials described the match as “absolutely incontrovertible”and a “bingo match”. As a former U.S. soldier, Mayfield’s fingerprint was on thenational fingerprint system. Mayfield was imprisoned for two weeks. The fingerprint,however, was not his. According to one law professor,“The Mayfield misidentification also reveals the danger thatextraneous knowledge might influence experts’ evaluations. If any ofthose FBI fingerprint examiners who confidently declared the matchalready knew that Mayfield was himself a convert to Islam who hadonce represented a convicted Taliban sympathizer in a child custodydispute, this knowledge may have subconsciously primed them to“see” the match. ... No matter how accurate fingerprint identificationturns out to be, it cannot be as perfect as they claim.” 435When Mayfield’s personal information was combined with the crime scene evidence,the FBI was convinced of his culpability. Yet according to a recent panel of experts,they were wrong. 436 As the collection of biometric information increases, and as itmoves from law enforcement to civilian applications, the error rate may significantlyincrease.Iris recognition and blind and visually impaired peopleIris recognition is a relatively new identification technique. In the decade since the irisidentification algorithms were patented, nearly all technical reports and trials have beenconducted at a general level. It appears that no trials have been undertaken with specificreference to blind or visually impaired users. When such people are unable to use asystem for whatever reason, they are referred to within the biometrics industry as the“outlier” population (the members of which are colloquially known by the industry as“goats”). 437 They are frequently excluded from research trials. The reported levels ofaccuracy and acceptability of iris recognition therefore tend to be based on analysis ofthose who are physically able to use the technology rather than representing a crosssectionof the community.A distinction should be made between the “outlier” population – those who physicallycannot use the technology – as opposed to the population who would find thetechnology difficult to use or who would produce inconsistent data. The latter group434 ‘The Achilles' Heel of Fingerprints’, J.L. Mnookin, Washington Post, May 29, 2004.435 Ibid.436 ‘FBI Faulted in Arrest of Ore. Lawyer’, B. Harden, Washington Post, November 16, 2004.437 See references to this term, for example, in http://www.speechtechmag.com/issues/3_3/cover/442-1.html

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!