America's robot army covers the growing number of robotic equipment being deployed or developed for use by the American Military.
The scenario being developed is where Robots replace infrantry troops, where networked surveillance equipment is able to track everything moving in whole cities, etc. Rather than place flesh and blood troops at risk, it will be the machines instead.
One thing that strikes me is this is an utter violation of Azimov's Rules of Robotics. Okay, Azimov was a science fiction writer, so maybe that doesn't give his opinions a lot of weight. But he was prescient enough to see the possibility of self directed robots. As computer power grows, so does the processing capabilities that can be built into a robot, and the greater degree of self autonomy can be programmed into that robot.
At what point does that robot become so autonomous that you can simply shout an order to the robot "Kill the enemy" and the robot is in charge of determining just who the enemy is? We humans have a tough enough time determining who the "enemy" is, that is determining which people on the battlefield should die or live. That's what "collateral damage" is, a failure of adequately determining the proper enemy.
What hope do we have of adequately programming a robot to do this well?
The article also discusses surveillance equipment of scary proportions. Under development is a whole range of small video devices, for example, that can be scattered throughout a city to blanket the city with surveillance. Once a city is blanketed with surveillance equipment, every movement is visible and trackable. Then, Human ID At A Distance is a project which would allow identification of people from that surveillance (HIAAD was part of the Total Information Awareness project).
The article discusses its use in war zones. Suppose it is deployed in the home grown war zone? American cities, that is.
Technology is technology, and it is a matter of the humans who deploy the technology to determine its use.