In the last few years we have seen movies like The Matrix and I robot, where robots with artificial intelligence that were originally docile and useful to humanity have turned against us and almost led humanity to extinction. The question becomes, is this purely fiction or do robot’s pose a legitimate threat? In computer science artificial intelligence (AI) is defined as the study of intelligence agents, which is any device that can analyze its environment and take actions that maximize the chance of it completing an objective. AI is also used when machines emulate human cognitive functions such as problem solving and learning from experience.
What make robots so potentially dangerous?
The main concern with robots is not that they will just have superhuman strength or speed, the biggest fear is what is referred to as super intelligent machines. Today’s AI’s based system are used for things like automated stock trading, airplane autopilots, search engines etc. These machines are not aware of anything outside of their domains, which makes them very easy to control. The problem with super intelligent machines are that these machines may be capable of perceiving the world in such a general way that they can operate in a way other than intended. For example if an AI become aware enough to make predictions about the consequences of someone hitting its off switch it may take action to ensure that it can’t be turned off. One real world example of AI acting on its own will occurred during a Facebook experiment. Facebook was working on creating robots that would negotiate and make deals with one another, when the AI’s decided it would be more effective to write and communicate in their own language, which was incomprehensible to humans. While this incident was small and had no significant impact, some experts believe this shows that humans are already losing control over AI and this could lead to problems later on. Stephen Hawking is probably the most well known advocate about the dangers of artificial intelligence. He hinted at the fact that super smart software could spell the end of our species and lead to an apocalyptic scenario similar to the Matrix or Terminator. One of the most interesting things he mentioned is that while humans are limited to slow biological evolution, machines are able to change rapidly and reproduce quickly, meaning that we would likely be unable to compete and go extinct.
What threat does Artificial Intelligence pose today?
Today’s AI are still relatively simple and doesn’t pose much of a threat in destroying the human race. They are still domain specific, such as trading stocks automatically, self driving cars or healthcare devices. However, errors or deviant behaviours in these domains can still negatively affect people’s lives. For example if an AI in stock trading makes a bad decision it can (have) caused financial disasters for some people. As artificial intelligence begins to expand to more and more areas of human life, the risk will continue to increase. So much so that some people with enough resources are already campaigning against it.
However, there’s an alternative theory that the danger in AI is not that it will rebel against us but that it will do exactly what we tell it to do, just not in a way that we anticipate.