Walk, don't run
- By Thomas R. Temin
- Sep 10, 2003
Thomas R. Temin
When local police throw out a technological toy, you can be pretty sure it's worse than useless. Cops have an exquisite sense of what can do them good and what can't. Thus the Tampa, Fla., police recently tossed out a facial recognition system that, after a two-year try, didn't nab one criminal.
Similarly, a report obtained and circulated by the American Civil Liberties Union showed that during tests at Logan International Airport in Boston, facial recognition systems failed to spot targeted people about 40 percent of the time. In that test, airport employees' mugs were put in a database, and those workers would go through camera-equipped security checkpoints.
Facial recognition, a still-maturing technology, may well work in one-at-a-time authentication applications. But in rapid scanning by distant cameras of large crowds coupled with continuous database searches'well, it just doesn't.
This technology misfire shows the perils of looking for instant technological solutions to complicated problems. Until the policy and management priorities are sorted out, technology won't help much.
Everyone is worried about the vulnerabilities posed by the country's creaking electric grid. Without juice, the cyber infrastructure isn't much good, either.
But the grid's susceptibility to failure stems as much from mixed-up public policy and utility management as from old technology. Fixing the first two will provide the road map'and the capital'to fix the third.
In the absence of clear objectives, technological so-called solutions can be expensive but ineffective. Logan Airport officials took a lot of heat after 9/11 because two of the doomed flights took off from there.
Even so, officials there would have been unwise to simply buy and install something as gee-whiz sounding as facial recognition. Patient testing proved this.
When the pressure is on, such as in high-profile disasters, is the worst time to rush into new technology.