Explosive Detecting Devices Face Off With Bomb Dogs

First time accepted submitter titan1070 writes "French scientist Dr. Spitzer and his colleagues have been working on a device that can sense faint traces of TNT and other explosives being smuggled into airports and other transportation methods. the hope for this device is that it will surpass the best bomb finder in the business, the sniffer dog. From the article: ' While researchers like Dr. Spitzer are making progress — and there are some vapor detectors on the market — when it comes to sensitivity and selectivity, dogs still reign supreme. “Dogs are awesome,” said Aimee Rose, a product sales director at the sensor manufacturer Flir Systems, which markets a line of explosives detectors called Fido. “They have by far the most developed ability to detect concealed threats,” she said. But dogs get distracted, cannot work around the clock and require expensive training and handling, Dr. Rose said, so there is a need for instruments.'"

Comments

In other words

[NoNonAlphaCharsHere]

"We can't use dogs to spy on everybody, everyplace, all the time".

Re:In other words

[hrvatska]

Turned out, the dogs were responding to very subtle cues from their handlers, rather than their own senses. Which renders them completely inappropriate for law-enforcement use.

Please Mod -1000: Utter Bullshit.

Dogs are the absolute best tool we have for the job. There's a reason we use dogs to hunt animals, guard animals, property, and people, track fugitives, search for survivors, bodies, drugs, and explosives, detect cancer or seizures, lead the blind, etc. They have incredible senses and are very intelligent.

Please link to proof of your "literally unacceptable percentage of false positives" for properly trained canines and handlers.

Seems to me that she wasn't saying anything about properly trained handler and dog teams, but about the likelihood that so many trainers have biases that lead to false positives that dogs cannot be relied upon. She said "the dogs were responding to very subtle cues from their handlers." I don't see anything in that post about well trained dogs paired with unbiased trainers. It is very well documented that handler bias frequently leads to false positives. For example, this article notes that sniffer dogs got it wrong four out of five times in 14,102 searches. This articleclaims that over a three year period only 44 percent of alerts by dogs led to the discovery of drugs or paraphernalia. A UC Davis study found that if handlers expected their dogs to find drugs they consistently found drugs, even when there weren't any. A little bit of searching will turn up plenty of other examples. In some cases defenders of using dogs claim that the high rate of false positives is due to drug residue being left in a vehicle or on a person. That the mere presence of someone carrying a substance the dog was trained to detect, like marijuana, in a vehicle hours earlier could result in a false positive. Medical marijuana is legal in 17 states and the District of Columbia. Which means that just transporting someone to legally obtain some marijuana for a medical condition could result in being searched and detained.

Re:In other words

[Anonymous Coward]

If having an anon call you out bothers you (I admit he was a bit obnoxious about it), you should include links in the original statement. Anyone could have said, "dogs are perfect" or "dogs are worthless" without any references.

It's worth noting that the article you link refers specifically to bad training, and does not suggest that all detection dogs have those same issues with being trained to take detection cues from their handlers. It goes on to suggest that this is a problem elsewhere, and that there's too little information (or apparently, oversight in training).

The original material, linked in the summary, suggests dogs are the best thing we have right now. That's from Flir, who sells the alternative devices. They even suggest that their sensing technology is meant to be used only as a complimentary test, along with dogs.

I just I don't think it's appropriate for us to start drawing firm conclusions on the abilities of dogs, trained and handled properly. It seems they're the best thing going, made better in combination with technology, and what we really need is to keep an eye on how they're trained.

Priorities....

[mark-t]

They have by far the most developed ability to detect concealed threats

That statement, entirely by itself, should qualify dogs as a better option, but let me elaborate...

But dogs get distracted, cannot work around the clock and require expensive training ...

so do employees. What's your point?

Dogs work. They work well. They are unsurpassed in reliability by any instrument we've been able to devise.... the fact that they can't be used like machines could should no more be a reason to not use them than the fact that humans can't work like machines should be a reason to not employ people.

When a machine can do a *BETTER* job at it than a dog... then I could see replacing them being viable. Until then, however, let Spot and Fido keep their jobs.