Henning Beck: You cannot make a direct comparison between the human brain and AI. The brain is always better in situations with little data, where there is uncertainty, and with human interactions. Computers, on the other hand, are better when you have a lot of data and the data situation is clear and measurable. Computers follow rules, whereas we can set new rules and also break them. We think interactively and in concepts, and we change things. So the human brain still has the upper hand in many areas.

Computers work on the basis of input, followed by processing and then output. In the case of the brain, processing and output are one and the same. Thoughts are not results that are similar to a computer’s output. They are the product of interactions between neurons. In theory, the number of thoughts or states that a human brain can have or sense in a single moment is gigantic; it exceeds the number of atoms in the universe by a huge factor. The brain has 80 billion neurons. It is not data-based and does not compute individual steps using ones and zeros. Instead, it works with concepts or categories.

As well as taking in information and passing it on, we understand what the information relates to and we abstract what we understand to relate it to other things. For example, we very quickly grasp the meaning of new words, such as Brexit or Teuro. We understand the concept and can use it after hearing a word just once, because we understand the idea behind it. Current computer systems cannot do this. While a computer may be able to play chess very well, it does not know what chess is. It simply applies rules and provides the best answer in a system because it has previously been told what the best answer is. But no computer knows what chess pieces are, what a pawn or a king is, or that chess is a game.

We generally use the word intelligence as a generic term for human thought, even though it is only one sub-area of our full range of mental abilities. Human thought is much more than this one aspect. We are also creative, emphatic and communicative. The most intelligent person in the world can do nothing with their intelligence if there is no-one to convey it to. And it is of no use to me if I cannot break rules.

If you drive a car, you need to follow the rules of the road. But if you are building a new car, you have to abandon set ways of thinking and even disregard rules. That is an essential step on the path to acquiring new perspectives. You need to question things, even if you ruffle some feathers in the process, or fall flat on your face. Most inventors, scientists and artists come a cropper at some point, but they learn from their mistakes. If you are not prepared to accept that, you risk becoming simplistic. Many people focus only on speed and efficiency, but machines have these features as well. It is the error in thinking, not the level of perfection, that distinguishes us from the uncreative machine.

In situations where we are performing routine activities, such as cycling, hiking or cooking, the brain goes walkabout. Regions of the brain that plan and hypothesise then become active. This kind of cerebral meandering or drifting away is extremely powerful if you have been thinking about a problem beforehand. The technical term for this is “mind wandering”. In the brain, it is associated with what is known as the default mode network. This kicks in when we are doing nothing. It brings together ideas and impressions. We also know that many people have good ideas in an aeroplane. This is a stress-free space, where people cannot use their mobile phones.

We shouldn’t believe that! At the moment, a lot of people, including Hollywood scriptwriters, like to conjure up this scenario. But we still don’t have the technology for it. While it is conceivable in theory that a machine could be more intelligent than a human being, today’s technology only covers a tiny area of our thought, namely the recognition of patterns. Also, there is hardly a single trend in biotechnology that is progressing exponentially.  AI is also an energy-guzzling process. The brain works with 20 watts. This is enough to cover our entire thinking ability. AI needs an incredible amount of energy to recognise a picture of a penguin from 10 million images. To solve the problem, it requires entire data centres that need to be kept cool. If we wanted to use AI to reproduce everything the human brain is capable of, we would need a vast number of nuclear power plants to provide the necessary energy, assuming it would even be possible that is.