An autonomous robot may have already killed people

There have been several documented reports that the assassination of an Iranian nuclear scientist was done by an machine gun with pattern recognition software smuggled in and set up by the Mossad.

I think the article is referring to Turkish Army UASs used in Libya. The Chinese also has at least one UAS in that theater but it was the Turkish drones that received the attention for their ability to take out Russian made air defense systems. They were also used effectively supporting Azeri forces in their recent fight against the Armenians.

US cruise missile generally fly to it target on it own,it have a map ,that can be updated on flight, Google Cruise Missile Tercom

The end of humanity will be on the horizon when machines can both design and build themselves, without human control. At that point, we are doomed.

No, no, quite some earlier…I still very well remember that 2016 shootout in Dallas with the first occurrence of a killing robot. Wikipedia also states ‘In the early hours of July 8, police killed Johnson with a bomb attached to a remote control bomb disposal robot. It was the first time U.S. law enforcement had used a robot to kill a suspect.’ Well, we still might dispute about the levels of autonomy of such robots 2016 and nowadays.

No, no, quite some earlier…I still very well remember that 2016 shootout in Dallas with the first occurrence of a killing robot. 

You are missing an important difference. The robot you are referring to had cameras on it and was being guided by a human operator. Humans were operating the robot through a communications channel and making decisions to place the bomb where it was placed. The robot didn't do any of that on its own. The same is true of, say, an MQ-9 Predator. Yes it's unpiloted, but it is providing an operator somewhere else a live video feed of the scene through sensors on the UAS via a satellite data link and humans have to make the decision where to fly the Predator, identify targets and make the decision to use a weapon. For the US at least there are almost always lawyers and at least mid grade officers making the decision to shoot. The Predator cannot do that on its own.

What the article is referring to are new unmanned systems that are programmed to think for themselves on a battlefield using AI, find targets and attack them without a human making the decision to use a weapon against a target. Basically turn them loose with no human in the link and let them kill what their computer programming and sensors think are enemy forces. If they make a tragic mistake, who is held responsible?

US cruise missile generally fly to it target on it own,it have a map ,that can be updated on flight, Google Cruise Missile Tercom

The choice of the original target and any change of target made in route are all choices made by humans. The weapon itself does not select targets independently.

This is great news, weapons systems could be programmed to eliminate redundant individuals, excessive consumers of scarce resources, non productive citizens, and other non participants in the world economy.

As long as the system is programmed to not discriminate based on race, creed, ethnicity, gender, religion, sexual orientation, age, or ideology, no one should complain!

Also known as autonomous loitering munitions which fly/hang around until a profile-fitting target is spotted.

1glennOct. 3  01:50 pm JST

The end of humanity will be on the horizon when machines can both design and build themselves, without human control. At that point, we are doomed.

A machine can never 'think' for itself. All computers and robots have to be programmed by people. We're not at the point of 'Terminator' yet.

WALL-E

Asimov’s rules never extended to the real world where people actively seek to destroy and kill others.