A sustainable and targeted Digitization strategy is therefore of central importance for companies in order to serve as a guideline in uncertain times. Even before the crisis, we realized that digitalization means change; now we know that change also requires digitalization. We asked founder Konrad Krafft whether and how current changes are driving digitalization and how digitalization brings about change.
After the outbreak of corona, it was discovered that a artificial intelligence had predicted the pandemic based on data. Does this show us that we should rely more on data-based decisions in the future?
A clear yes and no. Humans have always been dependent on information. Thousands of years ago, people needed to know which plants were edible and which were not. They passed this knowledge down through generations. A decision to eat the wrong plant could be fatal in case of doubt - in the best case, it only caused stomach ache. We make Decisions are always made on the basis of available information - consciously and unconsciously. Processing information and data is therefore nothing new.
So when computers make decisions, should we ask just as critically what data they are based on and what data the AI has been trained with?
A "new" addition is the Digital processing of information. Information processing works at the speed of light and a machine can take infinitely more information into account than a human being.
So the machine depends on the human performance of information processing?
In a human being, the "Memory" limited. It can juggle a maximum of five to seven factors in a conscious decision - in a gut decision, there are up to 20. We can perform around 1013 analog computing operations per second. A machine can already do much more today. The BlueGene/L supercomputer from IBM, for example, can perform up to 3.6,1014 floating point operations per second with twice the accuracy. 1
The machine therefore has the Speed advantage including the Amount of datathat it can process. However, it also has two decisive disadvantages:
- Unlike the human brain, it cannot collect information itself. We have to "feed" the machines and train them (deep learning).
- The human brain is highly interconnected, resulting in massively parallel processing that computers are not yet capable of. Nevertheless, chip designers are working on eliminating this disadvantage. 2
The Roboadviser from Minveo AG, for example, has landed an absolute lucky strike in the area of shares in real-time trading during the coronavirus crisis. Based on AI, market risks were recognized in good time and shares were sold within seconds. 3 As you can see, technology speeds everything up and allows much more in less time. Those who are fast and have enough data can make more profits - is the simplified formula.
So decisions with the help of AI can take us further?
Definitely. But we have to keep in mind that decisions can be just as wrong - just like in humans. Large quantities correct data are the decisive factor in the field of AI. A bias due to insufficient data can Wrong decisions favor - in technical jargon this is called "Bias effect". The assumption that computers and AI are infallible is wrong, as they are "fed" or programmed by us. As much as decisions in real time - without human intervention - are advantageous, they are in many places inaccurate. Ethically difficult. This applies not only to obvious cases in the military, where it is a matter of life or death, but also in securities trading. If incalculable damage could occur, an AI decision should not lead directly to an action, but should be confirmed by a human. Especially when a machine does not have a complete picture of the world, which includes moral principles. I therefore primarily see the AI as an assistance systemthat suggests decisions to us. In the long term, however, we will not be able to avoid giving the machines Rules for morality to teach.
So we need ethical principles for machines?
I am a fan of transferring guidelines that have been developed in the real world to the digital world. For example, who is liable for a wrong decision? A Control mechanism is important. We have a decisive advantage with machines, which I would like to emphasize as an opportunity: by having a complete data basis, we can Decisions tracked transparently be analyzed. In other words, why the action was taken or what data was "fed" into it. This is not possible with a human being, as it would contradict the principle of protecting the individual person. Data for the development of AI must therefore be carefully selected. The quality and thus the degree of accuracy of an AI is determined by the selection of data - The better the data, the better the decisions.
In order to safeguard good decisions, there may one day be a KI TÜV for machines. Just like cars, which are not allowed to drive on public roads if they are not regularly inspected.
What are good data-based decisions?
This is difficult to answer in general terms. A "good" decision should always have a Added value bring, it should Moving things forward. You could also say that, from a corporate perspective, the best decisions are those that achieve the best sustainable Cost-benefit ratio can help here. Data can help here, as patterns from the past that have led to a good cost-benefit ratio can be recognized if a good data basis is available. The task for an AI is the same as for a human being, namely to search for these patterns.
What is your forecast for the future?
Faster and better decisions with the help of data and AI technologies will come.
The advantages are obvious. When we use customer usage data to new products this offers enormous potential for Value creation potential. Or when, based on data Repair devices in good timevaluable Resources spared.
But I think that the question of the Quality of software-based decisions will continue to take center stage. They can provide the speed, we can already see that today, but the holistic quality of these decisions will be the subject of developments in the coming years.
Sources Text:
1 https://de.wikipedia.org/wiki/Gehirn#Leistung_des_Gehirns ^
2 https://m.heise.de/tr/artikel/Hardware-mit-Hirn-1478384.html?seite=all ^
