AI in management: “AI can make predictions but not decisions”

Enlarge

Professor Stefan Feuerriegel has been head of the new Institute of Artificial Intelligence (AI) in Management at LMU Munich School of Management since August 2021. Being also affiliated with the Faculty of Mathematics, Computer Science and Statistics, he researches the use of AI in business, public organizations, and healthcare. His latest article entitled “Bringing Artificial Intelligence to Business Management“ is published in the research journal Nature Machine Intelligence.

What tasks can managers leave AI to handle?

Stefan Feuerriegel: AI can take on clearly defined, repetitive, and less complex tasks. Or, according to the one-second rule, everything that humans could do themselves in one second. For example, AI can extract the invoice total or the invoice number from a document and automatically instruct a bank transfer — which, with a million customers, makes things highly efficient. But whenever emotionality, creativity, and cognitive thinking are required, human input is needed. In a planned corporate merger, for example, AI can supply information on how similar one company is to another — as long as the relevant attributes can be broken down into numbers. But whether that company is pursuing the same strategy as the other, or whether it’s even a good idea to buy out a company with the same focus, that takes a manager’s decision. AI can make predictions but not decisions. That’s why I don’t believe that AI will soon be doing all of our jobs, as some people fear.

Touchpoints with AI

So, which areas of a business can AI be used in?

Wherever large amounts of data are being generated that can be used to help AI evolve. In Marketing, for example, where companies send thousands of promotional emails or track website visits. Departments that process financial data are also well suited: In Accounting or Controlling, AI can help work out how much liquidity a company has.

But whether to put that money into acquiring a company, into research and development, or whether it should better go towards interest repayments — these more complex decisions need to be made by a human. Where AI cannot be used is in questions of strategy: How can I win new customers? What makes a successful product design? AI can only provide input here. AI normally lacks the data points for complex decisions — and when humans are selecting which data to provide, they are already making decisions, albeit subconsciously.

To what extent is AI already being used in companies?

Many companies in this country already have some touchpoints with AI and are trying it out on one or two projects. But full implementation is usually lacking, as is a comprehensive AI mindset among the workforce.

Businesses in Germany, and in Europe in general, are lagging behind those in the United States and parts of Asia. This is partly due to the fact that there is no ready-made AI business product, to put it bluntly, that managers can open, like Word or Excel, with a few clicks.

AI applications first have to be tailor-made for companies, which is why graduates with AI expertise are desperately sought after.

Another major obstacle is that many processes in German companies and institutions are still paper based. There just isn’t the digital data with which to use AI.
In order to deliver an organizational change towards artificial intelligence, a great deal of support is needed from management level, and absolutely all employees need to be given training in it.

Many think AI is a kind of all-purpose tool

What risks do you see in the use of AI in businesses?

Especially in Human Resources, companies need to be very careful because the risks around data privacy and security are simply too great here. All the more so in the public sector and in places where sensitive data such as health profiles are involved.

You can’t use AI with blind faith, you need to apply appropriate caution. In Germany, companies are obligated to comply with ethical standards when using machine learning anyway. Generally speaking, awareness of the risks of AI is very high here.

Do managers today know enough about AI?

Not really, because it is a brand new, very complex, and mathematical technology. Many managers think AI is a kind of all-purpose tool, a “Swiss army knife” that can do anything. It isn’t. Where AI works and where it doesn’t is certainly easier to understand from a mathematics or computer science perspective than from a business perspective. That’s why it is so important to advance this knowledge in Germany and Europe through teaching and continuing professional development.

Managers don’t need to know the AI’s algorithms, but they do need to understand certain basic rules for how to use them.

How do you use artificial intelligence in your research?

In our institute, we develop new AI management applications, roll them out in companies, and evaluate their impact on society: How should development aid be allocated to have the biggest impact? How can (anonymized) cell phone data be used to predict the future course of the Covid-19 pandemic? How can AI improve product quality to reduce waste in production? We address problems like these — that have a direct impact on sustainability — with new AI that we create ourselves.

Publication:

Stefan Feuerriegel, Yash Raj Shrestha, Georg von Krogh and Ce Zhang: Bringing artificial intelligence to business management. In: Nature Machine Intelligence