Pre-crime in Public Has Arrived - How Much More Can We Expect From Government
Posted by Zenphamy 11 years ago to Government
"Meet AISight – The Artificial Intelligence Software Being Installed on CCTV Networks Globally
Posted on April 24, 2014
If you thought that CCTV cameras tracking your every move in public was bad enough, you’re going to just love AISight (pronounced “eyesight” of course). The invention of a Houston, Texas based company called BRS Labs (which stands for Behavioral Recognition Systems) is headed by former secret service special agent John Frazzini, and this Orwellian surveillance platform brings artificial intelligence to all of those creepy cameras that have been installed everywhere around you.
Apparently, this system is currently being installed in Boston, and has already been implemented in Chicago and Washington. In the event you live in these cities, I bet you’ve never heard of AISight, and more importantly, I bet there’s been little to no public debate.
The most disturbing part about this platform is that this artificial intelligence defines what is “normal” behavior and anything that falls outside of that narrow band can be flagged for “pre crime” potential."
Posted on April 24, 2014
If you thought that CCTV cameras tracking your every move in public was bad enough, you’re going to just love AISight (pronounced “eyesight” of course). The invention of a Houston, Texas based company called BRS Labs (which stands for Behavioral Recognition Systems) is headed by former secret service special agent John Frazzini, and this Orwellian surveillance platform brings artificial intelligence to all of those creepy cameras that have been installed everywhere around you.
Apparently, this system is currently being installed in Boston, and has already been implemented in Chicago and Washington. In the event you live in these cities, I bet you’ve never heard of AISight, and more importantly, I bet there’s been little to no public debate.
The most disturbing part about this platform is that this artificial intelligence defines what is “normal” behavior and anything that falls outside of that narrow band can be flagged for “pre crime” potential."
The scary part about this is how government will use it, but even that is pretty limited. Barring how often police already abuse their power, theoretically you can't arrest someone from behaving suspiciously, as it should be. At best it would simply be a tool used to keep police nearby and at the ready in case the potential turns actual. Of course police will abuse this system to harass innocent people, but they already do that anyway.
But it's notable that in the article, it's police departments applying the software to existing public CCTV. If you give the police something based on history - they will abuse it.
Example:
People walking through a metal detector in front of a hotel = normal, no alert
People walking AROUND the metal detector = abnormal, alert raised
If a the computer sees a certain pattern of behavior on a regular basis, then that behavior is classified as "normal" and gets ignored. Only things which are out of the ordinary (i.e. things it doesn't see very often), will trigger an alert. And even then the alerts will still be reviewed by actual human beings to determine whether or not it was a legitimate alarm.
That said, I do see a potential problem in which the system might inadvertently flag ethnic minorities if they happen to enter a predominantly white neighborhood, as the system might recognize dark skin as being "abnormal" for a particular scene, since that camera might not see dark skinned people very often. But that's something the computer can be programmed to ignore, so it shouldn't be too hard to fix.
Is this type of system something that will need to be fine-tuned? Absolutely. But it still provides a more efficient method of protecting large installations.