St. Gallen Executive School’s Johannes Binswanger provides a realistic assessment of the impact of Big Data and AI on how we work
Over the past few years there has been considerable hype around the coming of AI to the corporate workplace and the potential power of Big Data. Much of it may be justified but much is still speculation.
In this salutary interview, for the University of St Gallen Executive School of Management Technology and Law, Jennifer Maier asks Prof. Dr. Johannes Binswanger a series of fundamental questions that get to the practical reality of AI and Big Data and their impact on the future of the way we work:
Prof. Dr. Binswanger, how much has Big Data and AI already changed our world?
Big Data encompasses a relatively small group of superstar companies. They include, for example, big tech companies like Amazon or Google and startups like Stripe or Revolut. These companies are putting other companies on the spot. What’s amazing here is how much Big Data and AI are shaping these companies, and thus our perception of a world full of Big Data. However, processes of about 80% of companies in the German-speaking world have hardly changed in their use of data-based value creation.
Why is it nevertheless important that today’s manager understand something about AI?
In many companies, data-based applications are still waiting to be discovered. For this potential to be discovered, managers need to get a sense of what data-based applications, which we often refer to as ‘AI’, can actually do. The prerequisite, of course, is that they know which business processes (could) leave behind which data traces and what this data looks like. This plays a decisive role in ensuring that managers have realistic expectations regarding applications and their benefits. Thus, when implementing and evaluating AI, they should be proactively involved in the process. Finally, executives also need to approve the necessary money for digital processes and integrate the projects into a strategic masterplan. ‘Strategic’ is crucial here, because a new focus on a data-driven value creation strategy includes a strong strategic component. Accordingly, executives should be integrated in selecting and implementing AI-based processes and understand something about the subject.
Will all executives need IT skills in the future?
IT skills and an understanding of what data-based solutions can achieve are fundamentally two different things. By IT skills I mean the use of IT tools. That is indeed becoming more important. However, executives can have good IT skills and still be clueless about data-driven value creation and AI. This would tend to reduce AI’s chances of success.
Will all managers need to know how to program in the future?
Understanding what data-based solutions can achieve has nothing to do with programming. Rather, it’s about an executive knowing the potential that data can offer to manage a machine park or warehouse more efficiently, to better meet customer needs, or better manage financial planning. You don’t have to know how to program to have this understanding. However, one should be aware of how such problems in the company can realistically be solved with data and where this would be hard to achieve. Based on this, executives should define an appropriate data strategy.
How will we deal with data in the future? Will we become transparent people for analysis purposes?
I hope that this scenario will not materialize. On the other hand, processes could be made more efficiently in many companies if we had ‘glass machines’ and ‘glass warehouses’ for production processes. In marketing, the risks of violating privacy standards are higher. However, a great deal of data already exists and is also used, but often sub-optimally. Using this data in a better way would not lead to any additional privacy issues. Who is not familiar with this scenario: advertising campaigns with a 10% response rate (if not less). This can be done in a better way using existing data.
How can Big Data help us deal with uncertainty?
Big Data fundamentally helps improve regularly recurring uncertainties. Supermarkets should not regularly be surprised if many people buy barbecue meat on sunny Saturdays (for the future of the climate, I hope it will soon be plant-based barbecue). We humans are bad at seeing through complex patterns with many influencing factors, even when the complexity results from regular patterns. This is where Big Data and algorithms come in to help us identify recurring patterns.
In addition, there are many instances of uncertainties where Big Data will do absolutely nothing: Will there be two new economic blocs in 2035, dominated by China and the West, respectively? Will we have come to grips with the financing of climate measures by 2035? Will Switzerland have found new regular relations with the EU? In the language of statistics behind all the algorithms, what we have here is a ‘nonstationary data-generating process’. If that’s the case, Big Data is of little to no help.
Do you recognize ethical or moral problems when dealing with data? To what extent can data also discriminate?
When I look at projects from participants in our Executive MBA program, the applications of data are often not critical. Regularly, it’s about inventory management and the machine park. The moral issues surface primarily when data-based solutions are used in sensitive areas. Consider, for example, the following data-based decisions: should prison inmates be released? Where should police operations take place? Should a person be monitored by the state because they might be a security risk? Often, these are government responsibilities. But the state often works with private providers here, especially in Anglo-Saxon countries. Data-based applications in the insurance sector can also be problematic. For example, if someone does not receive supplementary insurance because the cell phone has detected a health risk based on a person’s gait and voice. The monitoring of employees for the purpose of increasing productivity can also be classified as problematic.
Where is this leading? What developments do you see in the future of Big Data and AI?
As already mentioned, there is a rather extreme segmentation in the use of data-based solutions. While the superstar companies are far ahead, probably more than 80% of German-speaking companies have yet to discover a first project that deserves the label ‘data-based’. Of course, it sounds cooler if you call it ‘AI’. I’m a bit surprised how many companies are still in sleep mode here. For many companies, AI is a priority for the next three years. The risk for those 80% of companies is that startups and superstar companies suddenly undermine their business. Then it’s too late to catch up.
What about the regulation of AI? What rules do we need to be able to use this technology successfully and without danger?
How regulations should be structured is a difficult question. Especially around the issues of security, insurance, and employee monitoring at private companies. The EU is in the vanguard of development here. It is just as difficult with topics that can influence political opinion-forming and social cohesion, for example via social networks. However, many young companies that want to make their processes more efficient by using data are not affected by these sensitive areas.
Jennifer Maier’s interview with Prof. Dr. Johannes Binswanger, 'Artificial Intelligence—Curse or Blessing?', was originally published on Vista, the blog of the University of St. Gallen’s Executive School of Management Technology and Law.
Our Executive Education programmes are characterised by its high level of relevance for current practical issues, by drawing on the latest results in research.