Artificial Intelligence (AI) – the topic of the 21st century. What is clear is that the use of AI in everyday life is advancing. But this inevitably raises the question of whether or how our professional lives will change. What kind of change do we have to prepare for and which areas are affected by change? The use of AI and machine learning is also becoming increasingly relevant in controlling.
According to the Gabler Wirtschaftslexikon, AI means “research into ‘intelligent’ problem-solving behavior as well as the creation of ‘intelligent’ computer systems. Artificial intelligence deals with methods that enable a computer to solve those tasks that, when solved by humans, require intelligence.” (Prof. Dr. Richard Lackes, Dr. Markus Siepermann, “Künstliche Intelligenz”, in: Gabler Wirtschaftslexikon.) The origin of the term goes back relatively far, to 1956, when computer scientist John McCarthy was looking for a title for the literature of a specialist conference at Dartmouth College in Hanover, New Hampshire (USA). Even earlier – in 1947 – Alan Turing addressed the question of whether machines can think at a symposium in Manchester.
Nowadays, AI is divided into “weak” and “strong”. Weak AI supports human cognitive processes in solving individual problems. The final decision the ultimate action lies with the person himself. This corresponds to the current state of the art. Strong AI, on the other hand, can solve complex issues. The cognitive performance can be compared to that of a human: for example, strong AI can make decisions.
Weak AI is already being integrated into everyday life through systems such as Siri or Alexa. Furthermore, the so-called Face ID for unlocking the smartphone or the algorithm at Facebook that creates individual news feeds for each user, for example, are considered AI. But the topic is not bypassing controlling either: Machine learning, robotic process automation, and natural language processing, for example, are under discussion. Initial attempts to use AI are generally in the direction of digitization and automation. The aim here is to enable controlling to work more efficiently. With the integration of AI, for example, visualization in standard reporting or exploratory analysis can be improved. AI in the form of predictive analytics can also be used to combine various internal as well as external information and to model different financial scenarios. AI can therefore provide support, particularly in forecasting or planning. This relief would allow controlling to focus more on those tasks that cannot be performed by machines due to the cognitive performance required. In particular, emotional intelligence, contextual knowledge or social skills cannot be replaced in any case.
Despite the potential to make work easier for people, a survey conducted by the Bundesverband Digitale Wirtschaft e.V. of 2018 with 1044 participants on the topic of AI yielded interesting results: According to the survey, 69% of respondents fear that AI will replace jobs. Furthermore, 48% are afraid that humans could lose control. 74% of respondents also feel that the human element is lost when machines make decisions. From these concerns, it is clear that the German population intuitively thinks of “strong” AI when it comes to the topic. Therefore, they are also very skeptical in this regard. This is also reflected in controlling. Confidence in AI is still lacking, as too many aspects have not yet been finally and satisfactorily clarified. For example, secure data storage and the question of liability are causing uncertainty. The traceability of results is also not sufficiently given with the use of AI. It is currently seen more as a kind of “black box”, since the controllers cannot recalculate the figures they get from predictive analytics. Instead, they must rely on the accuracy of the results in the situation. For this reason, AI is not yet trusted by management. Difficult and at the same time very important disciplines – such as forecasting and planning – are reluctantly left to a machine.
In the future, the use of AI in controlling would be associated with many challenges. In addition to a reorientation of the corporate culture and the professional image of the controller, new solutions must be integrated and accepted. To achieve this, automation would first have to make further progress. Subsequently, AI could be tested for specific use cases. External factors such as market growth can be calculated well using predictive analytics. The results can then be easily incorporated into the forecast using value driver trees, for example. This could then be followed by the implementation of AI in other projects.
Conclusion – AI in Controlling
Overall, it is clear that weak AI can help controllers do their jobs. However, a long transformation period is needed before strong AI can be implemented. There are too many questions in this regard that have not been clarified, creating uncertainty among the population. Accompanying changes must be well planned and executed to reap benefits from AI. The position of the controller will therefore remain important in the future, and the use of AI will be less of a threat and more of a support for him.