Sorry, but no. They're covering miles of outdoor perimeter. The problem is that there are very few technologies that can reliably detect things like people, while ignoring things like various animals (or heat from jet engines, etc.).
(1)I think you could buy them commercially at least as early as the 70's
In fact, I rarely hear of new deployments considering these technologies (though I'm sure there are plenty of them, as the companies are still in business).
If a 100MM taxpayer funded system can't tell the difference between a jet engine heat output and a person's heat signature, then someone should be criminally liable for defrauding us.
And as far as differentiating a deer from a human, it should have detected the moving object, sent the images to a person sitting in front of a computer, and he/she could have classified the object.
JFK is a complex airport, I'm not aware of the specific length of the perimeter they are covering there, but it's likely between 8 and 15 miles of perimeter fence line. You need something with a high degree of reliability and false alarm immunity to make this work. This is especially true because the fact is that the probability of a real terrorist incursion is pretty low. So, the guards aren't going to stay actively engaged with reviewing alarms for very long if essentially 99.9% of what they see are all false alarm/no-action events.
Video analytics can be very reliable, but you can't rely on humans to weed out the false positives very effectively. I've heard of (heard of, as in been involved in replacing them) video analytics systems that generated hundreds of false alarms per night on similar sized deployments. Any human operator is going to quickly become immune to the alerts when they happen at that rate.