For the “mindless/blind use of a database may lead to excessive results” file, from a wire service story:
Peace activists Medea Benjamin and Ann Wright have been arrested in the U.S. while protesting the Iraq war, but they never dreamed that would prevent them from entering Canada.
The arrests landed Benjamin’s and Wright’s names in an FBI-run database, the National Crime Information Center, which Canada also relies on to screen visitors. When the two women visited the country in August, they were told they would have to apply for “criminal rehabilitation” and pay $200 if they wanted to visit again. Neither did.[...]
Now, Benjamin and Wright are asking why the names of people arrested during peaceful protests would be included in an FBI-maintained database meant to track fugitives, potential terrorists, missing persons and violent felons.[...]
Canada generally refuses entry to anyone who has been convicted of a criminal offense, regardless of the nature of it, [a Canada Border Services Agency spokesperson] said. Those with convictions, however, may apply to be rehabilitated, which involves filing paperwork and paying a processing fee ranging from $200 to $1,000.
One would think that the is (or ought to be) a few codes and/or flags in the database to provide some indication as to the nature of one’s presence in the NCIC database, and that those codes could be used to tweak the filter, or at least prompt a border agent to ask a few questions to clarify the situation.
Admittedly, I think it would be understandable that Canadian authorities might want to keep known rioters out of their country, and that a simple database may make it hard to differentiate between the fostering of anarchy and being detained for exercising one’s freedom of speech. But still, the rise of the database, the adoption of predictive modeling, and the expansion of data-mining and algorithmic decisioning shouldn’t preclude some good, old-fashioned common sense.