Vviewpoints
DOI: 10.1145/1400214.1400223
Privacy and security
A Multidimensional
Problem

It’s not just science or engineering that will be needed to address security concerns, but law, economics, anthropology, and more.

WhEn thosE oF us who are now editors of this magazine were in graduate school, it was easy to believe that with the inevitable exception of automation, social implications of computing technology could be ignored with impunity. Yes, before the public Internet, there was discussion of social impact—Joe Weizenbaum’s ELIZA, the Department of Health, Education, and Welfare report, Records, Computers, and the Rights of Citizens,^a the establishment in Europe and Canada of data commissioners, the “I am a [person]. Please don’t bend, fold, spindle or mutilate me,” joke that made the rounds in the 1970s,^b the role of computers in President Reagan’s Star Wars program—but it was easy to maintain the fiction that the machines we built and the code we wrote had as much social impact as the designers of John Deere tractors have on the migratory patterns of cliff swallows: minimal and not really significant.

Tom Lehrer once sarcastically characterized a famous astronautics engineer, “‘Once the rockets are up, who cares where they come down? That’s

 

illustration by phil hulinG

a This report, which recommended legislation supporting Fair Information Practices for automated data systems, was highly influential in both Europe and the United States; see http://aspe.hhs.gov/DATACNCL/1973privacy/ tocprefacemembers.htm.

b This was a takeoff on IBM’s instructions for the handling of punch cards.

not my department,’ says Wernher von Braun.” ³ But while the view that scientists bear responsibility for the social impact of their work was perhaps radical when it was espoused by Joseph Rot-blat (a nuclear physicist who later won a Nobel Peace Prize for his work on nuclear disarmament) in the decade after Hiroshima and Nagasaki, this expectation is no longer unusual. It is also no less true for technologists now than for scientists.

This is part of the ACM code. The original ACM Code of Ethics and Professional Conduct stated, “An ACM member shall consider the health, privacy and general welfare of the public in the performance of the mem-

ber’s work.” It went on to say that, “An ACM member, when dealing with data concerning individuals, shall always consider the principle of individual privacy and seek the following: To minimize the data collection; To limit authorized access to the data; To provide proper security for the data; To determine the required retention period of the data; To ensure proper disposal of the data.” (The current ACM code of ethics contains a similar set of principles, though it omits the requirement regarding proper disposal of data.) But observing current computer privacy and security practices leads one to question whether this code is honored more often in the breach.

Each week brings yet another news story of a major security breach, of the ability to do a cross-site scripting attack on the new favorite mailer, of the polymorphic virus code that changes its signature to evade detection. We aren’t getting privacy and security right.

We aren’t even asking the right questions. A recent U.S. Department of Defense (DoD) effort to develop an Iraqi ID database of bad guys is one such example. The database includes not just names, but biometric identifiers: fingerprint records and iris scans; its purpose is to maintain records on the people who keep turning up in an area soon after an explosion has occurred. ² As any developer knows, of course, this database will not be used only in this way. One such likely use will be at checkpoints—and currently in Iraq, it can be