Killer Machines Controlled by Artificial Intelligence Artificial intelligence has been advancing in leaps and bounds. I believe it is just getting too smart. I say this, because it is at the point where people are overestimating its ability, but in a couple of years it will be far ahead of those estimates. What, you haven’t heard this before? That is because no one wants to talk about this, they would rather talk about the great advances in weapons, which rely on artificial intelligence. The militaries of the world want to put artificial intelligence in charge of areas which might lead to life and death decisions, without having a human available to check that decision. How would you like to be on an operating table and there is a slight glitch and a robotic doctor makes the wrong cut or goes a little too deep and severs your spinal cord or even nicks it causing you to become a paraplegic? You really don’t have to answer this question, because we all know what that answer would be. There is one thing we have to remember, a device even if it is based on artificial intelligence is still a device, and there has never been a device created which was perfect. Eventually all devices have a problem, even though it might take longer for some to have these problems to appear than others. What is making artificial intelligence even more scary is the fact scientists used to teach it by using human experiences. They would go back many years incorporating decisions made by humans over long periods of time and the AI (artificial intelligence) would learn this way, but this is becoming a thing of the past. More and more scientists, are beginning to think this is a waste of time and are skipping this method altogether. Some AIs are now learning by studying themselves, eliminating the human factor altogether. Game intelligence is learning by playing a game after studying its rules. It will play itself and then analyze the moves and keep doing this until it feels it has perfected what it is doing and becomes unbeatable. No one may worry about this very much when it comes to games, but what about other scenarios? Take war games for example. To the AI a war game is no different than a game of chess. It has no feelings for the amount of lives which might be lost. What might be the best solution to an AI might be one which costs far more lives than another. I know I wouldn’t want some AI planning an attack if I were still in the military. One of the scariest things is AI controlled weapons. There are so many reasons why this is wrong if there is no human at the controls. AIs only can refer to a complex set of rules when determining who is the enemy. It is much easier for a human to be completely sure who the enemy is. I read something the other day about how changing one character in a definition of what a certain animal was made it an AI think the object was a flower. Have you ever had your computer respond when trying to load something, that a device or software was corrupted? It doesn’t happen often, but when it does it is hard to correct. What if something like this happened to an AI controlled weapon, would it stop functioning or would it just kill the first thing it came into contact with? We have all seen how easily everything is being hacked. Just a year or so ago an American warship had its electronics turned off after a Russian plane carrying a strange pod under it flew near the ship. There is certainly no guarantee our AI weapons wouldn’t be turned against us by a hacker. I don’t know how many remember this, but years ago our missiles were fooled by the Iranians. They had pulled barges behind their ships and lit them on fire fooling our heat seeking missiles. There is probably nothing we could create which has a one hundred percent certainty of being hack proof. Rest assured as other countries are working on ways to hack us, we are also working on ways to hack everything they have. AI has the possibility of being dangerous and there is no getting around this. AI controlled machines were happily humming along a few years ago building cars. I believe all bulk builders of cars use them. They are supposed to be safe and shut down when a worker gets too close, and yet one machine in Japan had crushed a worker to death. The robot AI device used it arm to trap a worker against the machine. The machine was one which cut gears. The worker was a repairman. Japan has well over 100,000 robot AI machines in service. Death by robot or AI controlled devices has been happening for decades. In 2015 a lady working for a parts manufacturer in Michigan became trapped by a robot and was crushed to death. Before that a man was trying to retrieve parts and was struck in the head by a robot machine which didn’t detect his presence. Many other accidents have happened. Some have caused injury and some death. AI has no ethics, it only knows what it is told and only remembers this until something goes wrong. You can tell a robot device that human life is sacred on one hand, and that it is alright to kill other humans who are the enemy. There is something unique going on now and it is a debate among scholars about what to do when an AI finally kills someone and it is not an accident. They are arguing about who will be held responsible. Will it be the AI, will it be the creators, programmers, or none or all of the above.
|
