Artificial Intelligence

Thinking is more complex than we had long believed. The best example of this is our intelligence – ever since the days of punch cards, we have dreamed of intelligent machines that we can communicate with, that think for us and solve our problems. Seamlessly, like C-3PO in Star Wars or Data in Star Trek. HAL 9000 from 2001? Well, at least it was so complex that it could develop psychological conflicts. This is in stark contrast to the first well-known "AI," ELIZA, which was supposed to solve our human conflicts. But it only asked counter-questions. 

As our digital assistants on the relevant devices remind us every day, artificial intelligence still has a long way to go. Not that we haven't tried. But ELIZA marks only the beginning of a long chain of major and minor successes – and always high, but disappointed, hopes. In the wake of the current digitalization, hopes are once again enormous: big data and deep learning are showing impressive successes even beyond the limited rules of chess and Go, making hair appointments, for example, and thus passing the Turing test.

The model for current AI is the brain – which is only fitting, since the model of the brain has always been based on the most advanced technology of the time over the centuries. Neural networks can be trained for specific tasks and they learn at breakneck speed. In cooperation with the non-profit Hertie Foundation, we show where artificial intelligence stands today and where all this could lead. 

Manuela Lenzen provides an initial overview:Think, machine!