Let the machines do. How intelligent is Artificial Intelligence

摘要

“Intelligent” systems are present around us. Such machines are not only tools in our hands, they are able to make decisions and perform actions directly in the real world or in the artificial worlds of the internet. But even in the latter case, their decisions have consequences to our life. Many of them still act as “assistance systems”, leaving the final decision to the human user. The trend goes to more autonomy of the machines, even in critical situations when humans become overloaded by complexity. Additionally, humans are more and more willing to accept the proposals of the machines. But is this technique mature enough to guide or even to replace human decision-making? Especially, perception appears to be a hard problem for technical equipment. How accurate, how safe can decisions be in the case of incomplete and unreliable data? The paper gives some overview about recent developments in Artificial Intelligence and Robotics, their capabilities and serious problems from a general point of view.