EPIC logo

Panel Discussion

Security or Surveillance? Technology's Impact After September 11

Sponsored by

The Electronic Privacy Information Center
The Privacy Foundation

October 22, 2001,
1:00 pm - 2:30 pm

National Press Club,
Holeman Lounge
529 14th St. NW, 13th Floor
Washington DC

 

Summary

On Monday, October 22, the Electronic Privacy Information Center (EPIC) and the Privacy Foundation sponsored a policy briefing at the National Press Club in Washington, DC to explore the implications of new systems for identification and tracking on personal privacy. Questions considered included the reliability of face recognition technology, the implications of national ID cards, and the potential for regulating future identification technology.

The event was moderated by Marc Rotenberg, Executive Director of EPIC and featured five speakers; John Woodward, Senior Policy Analyst, RAND Corporation; Richard Smith, Chief Technology Officer, Privacy Foundation; Robert Ellis Smith, Editor, Privacy Journal; Whitfield Diffie, Distinguished Engineer, Sun Microsystems; and Jeff Rosen, Law Professor, George Washington law school. Larry Ellison, Chairman of Oracle Corporation, who has now become a central player in the push for a national ID card, was invited to speak at the event but did not attend. However, a five minute audio track, taken from the Oracle website, of his arguments in favor of such a card was played and an image of him placed on the speaker's podium! "We have been so busy protecting ourselves against our government, we have made it impossible for our government to protect us", Ellison said. "If we don't like our government, we just throw them out of office. [But we must give law enforcement] the tools -- like databases and ID cards -- and the latitude to protect us. And if we do, our liberties and our lives will be saved together."

John Woodward, spoke on recent developments in biometrics and facial recognition technology. He noted several difficulties with current face recognition systems. Most significant among them were the technological problems associated with capturing and analysing the data, and the intelligence problem of populating the database of suspected "bad guys". He stressed the importance of paying attention to the details of any system and evaluating on its own specific criteria. He discussed a possible example of a "FaceCheck" system, used to identify potential terrorists at the time they have to apply for a visa. The system would match a visa applicant's image against a database of potential terrorists. If there is a positive match it would be followed by a visual inspection by an officer. However, the system is not fool-proof and he pointed out several ways a dedicated terrorist might get around such checks. He argued that the system might nevertheless "raise the bar" for terrorists. The issue paper on which his comments were based is available free (in hard copy and electronic copy) from RAND.

Richard Smith CTO of the Privacy Foundation focused his comments on a technical evaluation of the leading face-recognition software called "FaceIt" made by Visionics. He demonstrated the weakenesses in the system by showing how it could not correctly correlate two pictures of a suspected terrorist. Factors such as lighting, background, and whether or not the subject is wearing sunglasses affect the reliability of the system. The reliability of the system is measured on a scale of 0 to 10 and it is important to set the right threshold value in order to prevent too many innocent people from being stopped, or letting too many guilty through. A casino for example may set a lower threshold to catch card counters, but that level would be unacceptable for government applications. Mr Smith's presentation is also available.

Robert Ellis Smith, an expert on the history of national ID cards, concentrated his comments on two main areas. The first was the issue of 'purpose specification' which is a central requirement of fair information processing and provides that personal information disclosure must have a definite purpose. He stressed that the problem with national ID cards is that they are multi-purpose and could therefore easily facilitate function creep, discrimination and tampering. The second issue he raised was that a national ID card requirement intrudes on personal privacy in such a way as to attack our "personhood", which he defined as the ability to act autonomously and be free to carry out the trivial tasks that make up the fabric of our lives, without outside scrutiny. Moreover, these cards would have to be handed out at birth to be effectively deployed.

Whitfield Diffie, was less skeptical of the potential for recoginition technology to improve dramatically and predicted that in the next 10 years we would become accustomed "ubiquitous recognition". Every storekeeper and government official would have access to a database with face scans to obtain the identities of individuals with whom they are interacting. He was also skeptical about the role of policy to regulate the multiple private and public uses of such technology.

Jeffery Rosen, emphasised the critical role of regulation in controlling the significant expansion of government power though technology. He referred to his experiences in surveying England's use of face recognition technology to show the wide potential for misuse of such a system (his recent article in the New York Times Magazine discusses this in greater detail). In practice low-level criminals rather than terrorists are entered into the database and the public has little control or knowledge over the information that is collected on them. He also discussed potential constitutional challenges to the use of new surveillance technology. The Kyllo majority opinion written by Justice Scalia last year does appear to protect privacy in the home but is also dependent on a reasonable expectation of privacy. The reasonableness, however, is constantly being undercut as cutting edge surveillance technology becomes more widely available. Congressional regulation and shoring up the Privacy Act of 1974 therefore emerge as the most important tools to prevent possible abuse of new identification technology.

Presentations were followed by a brief question and answer period and concluding remarks from the Chair.