What is Artificial Intelligence?

Artificial Intelligence (AI) is an indispensable part of our lives today and is already integrated in various applications. AI is one of the most promising technologies of the future.

Artificial Intelligence (AI) is an indispensable part of our lives today and is already integrated in various applications. AI is one of the most promising technologies of the future. As a new general purpose technology, AI has already been compared to the invention of electricity and the Internet. Kai-Fu Lee, scientist and former CEO of Google China, even writes in his book AI Super Powers: “AI is becoming more groundbreaking than the invention of electricity”1. But what is AI and what are the different types and application areas of AI?

Weak, strong and super AI

The terminology “Artificial Intelligence” was first used in 1956 by the US-American computer scientist John McCarthy. As a field of science, it deals with the research and development of intelligent computer systems that can solve problems and cognitive tasks that normally require human assistance. Typically, a distinction is made between weak, strong and super AI. Most of today’s AI systems can be classified as weak AI, because they can solve only a single and very specific task, even though at a very sophisticated level. Sometimes they even outperform humans. An example of this is the use of AI to evaluate computer tomographic images in medicine2. In contrast, a strong AI could solve many different tasks and would be therefore to some extent comparable to human intelligence. In the field of natural language processing (NLP), advanced language models based on transfer learning could be seen as an early, first step towards strong AI, since they can be used so solve a wide range of problems. Super AI, in turn, is the most polarizing technology, as it would outperform humans in all cognitive areas and applications. Thus, in the future, super AI could perhaps support humans in solving extremely complex problems and even develop innovations by itself.

Image1: Illustration of Artificial Intelligence

Machine Learning

In our everyday life we constantly encounter weak AI systems, which mainly occur in the form of machine learning algorithms in various applications. In recent years, they have become widespread, especially as “smart” functions in smartphones and in the context of other assistance systems. These sometimes very complex algorithms can solve well defined questions and problems, if they have been properly trained before. The algorithms autonomously derive correlations, rules and patterns from the training data presented to them in order to apply the extracted knowledge to new but similar situations and problems. Human support is only necessary to develop the basic technical framework of the AI. This independent learning of solutions for given tasks is therefore called machine learning. Due to the efficient and autonomous processing, trained AI systems can evaluate large amounts of data and gain information from it within a very short time. Machine learning systems almost always exceed human work in terms of speed and sometimes also in the quality of the results.

Application areas of AI

Due to the high efficiency, AI is already used in numerous different areas. As face recognition software in social networks and photo collections, virtual assistants on cell phones and in the living room, chatbots that answer customer queries or personalized advertising that is based on the user’s interest and usage behavior. AI is also used in more sensitive areas, such as road traffic, in the form of assistance systems in cars, which also enable autonomous or semi-autonomous driving. In the credit and banking system, AI is also used to detect fraud or to determine the creditworthiness of individuals and companies3. A further area of application is in medicine. There, AI is not only used for administrative tasks, but especially to detect and diagnose diseases4.

Image2: AI-based face recognition in real life

Author of this article is Robert Dehghan.