Artificial intelligence (AI) changes reality and, to an even greater extent, our view of it. Initially, a lot of fantasy and often very little actual knowledge combined to shape the fear that machines want to control the world. This then gave way to a fundamental fear that AI is to make people redundant in many sectors. Why is that the case? Where is the truth in it? In the digital age – in the times of machine learning and deep learning – how real is the vision that the "thinking machine" can simplify and improve our lives? In the following three blog posts we will take a closer look at artificial intelligence, starting – like most people do – with the fiction.
Let us recall: In 2029 machines will take over our planet, controlled by the Skynet computer system. The last people on Earth will fight fiercely for their survival. That is why a cyborg is sent back to 1984 to both terminate the future rebel leaders and kick-start the movie career of Arnold Schwarzenegger. Only one of these tasks was successfully completed, but ever since then, the world we know has been living in latent fear of clever machines and artificial intelligence.
Movie buffs will acknowledge this fear was being stoked far earlier by the likes of HAL-9000 in 2001 (1968) and Yul Brynner in Futureworld (1976), and was later continued by the epic Matrix movies (1999 onward) or sugar-coated and confused by A.I. (2001).
Reality and movies continue to develop hand-in-hand
Thanks to incredible technological capabilities and ever more potent computer performance, artificial intelligence has become a hot topic in many quarters. It goes without saying that it is a key focus for technology companies such as Apple, Facebook or Google, whose neural network recently beat a professional player in the board game Go for the very first time. The achievement caused quite a stir, since Go is a game based heavily on intuition and the outcome cannot be predicted through calculation alone.
At the same time, the plots of new movies such as Transcendence (2014) and Ex Machina (2015) are reaching a new, ever increasingly philosophical level. They no longer just process the human fear of being replaced or wiped out, but have an artificial super intelligence with feelings and objectives that could learn to manipulate humans and evade their control – a fear expressed very recently by the philosopher Nick Bostrom and previously by prominent admonishers of the closest thing we have to "real science fiction" – the world of technology, science and vision – such as Bill Gates, Stephen Hawking and Elon Musk.
Demise of humanity? Can it be reduced?
According to the history of cinema and popular science, our machines are getting more and more beautiful, rapid and clever. And this is true: Artificial intelligence has only just begun and holds enormous potential. If we want to exploit it, we need to stop fearing the advantages learning machines have to offer.
Machines are superior to humans when it comes to continuously carrying out recurring tasks – regardless of in which form or in which clearly defined environment, and regardless of whether it concerns industrial production, real-time bidding, stock trading, games of Go or, in the future, driving cars. They do not get tired, they carry out their tasks at consistently high levels of quality and do not get sick. There is a reason we have been building them for many years and are continuing to do so.
However: Machines are stupid and will remain stupid, even when it comes to artificial intelligence. Yes, they can learn more and more. Yes, they can recognize relationships and separate them from other situations. But they cannot develop their own theories, and they cannot identify and take on new challenges. In this respect, they are never superior to humans.
Artificial intelligence does harbor risks, but the task in hand is to overcome them in order to advance developments. Apocalyptic fantasies are misguided and, as the name suggests, pure fiction.