How to manage AI WISEly
Artificial Intelligence (AI) is rapidly transforming our workplaces. But what does this mean for managers and professionals aiming to leverage these technologies? How can you ensure that AI becomes not just an innovation, but a valuable and well-integrated solution within your organization?
Author: Edo Groot
AI systems can offer organisations many opportunities and fundamentally and positively change work in many ways. Marleen Huysman and Marlous Agterberg, researchers at VU University Amsterdam, together with Lauren Waardenburg (researcher at ESSEC Business School), describe in the book Managing AI WISEly how eight large and mainly Dutch organisations deal with implementing AI. Lauren Waardenburg: ‘We deliberately looked at what happens when large traditional organisations, so not the Google or bol.com of this world, but, for example, ABN Amro, Air France KLM or Centraal Beheer start using AI. What do they need for this? What needs to change? What do they need to focus on if they want to integrate AI into the organisation?’
Among other things, the research takes a closer look at the implementation of a fraud detection tool at ABN Amro, a chatbot in the customer service at Centraal beheer and an AI tool at the radiology department of Leiden University Medical Centre that predicts tumours. What is unique about this project is that the research focuses on tools that have actually been developed and implemented at organisations, and took a long-term look at how they work and it’s consequences.
7
Managing AI WISEly
Together with co-authors Marleen Huysman and Marlous Agterberg, Waardenburg presents four recommendations for W.I.S.E. management of AI, which requires work-related insights, interdisciplinary knowledge, sociotechnical change processes and ethical awareness. ‘That starts with organising for data; before you start developing the model at all, you need to have the data in place,’ he says. ‘What we often see happening at organisations is that people think: ‘we have a lot of data and we can probably do something with that. But in practice, it requires a lot more work. You have to generate incredibly large datasets that are also suitable for developing AI.’
AI systems vs. day-to-day work Once the data is in place, the model can be developed and the organisation can test and validate the model. ‘That is not just testing whether it is accurate, because that is often the idea: if we have a model that is accurate 90% of the time, that is enough. But the model must also fit within laws, regulations and organisational structures.’ However, this does not make a model immediately applicable to employees. Waardenburg talks about a gap that exists in many cases between AI systems and day-to-day work: ‘Think of the police, for example, where officers eventually have to start using a mathematical system to determine where to go to fight crime. Well, there's quite a gap there. Because police officers feel most comfortable on the street, rather than analysing behind a computer,’ Waardenburg explains.
Marleen Huysman, professor of Knowledge and Organisation at Vrije Universiteit Amsterdam: ‘there are a number of aspects that often go wrong when implementing AI in organisations. One is that organisations see AI as ‘yet another tool’, as something that belongs to the IT department. But by keeping it in one place you isolate it, whereas what you need is information from across the organisation. It is interdisciplinary, so it is important that everyone feels involved and has an opinion on it.’
'The more intelligent systems become, the more important humans become.'
Ethical challenges
Waardenburg describes that ethical challenges always arise when applying AI. For example, AI algorithms may continue existing discrimination or biases in the historical data, based on bias in application procedures so that certain groups are treated unfairly. Huysman: ‘as a decision-maker, you have to make sure you are involved all the way from the very beginning, in organising the data. After all, the data that is used determines what the system sets expectations on.’
Education for professionals
Marleen Huysman emphasizes that ‘the more intelligent systems become, the more important humans become. Managers need to understand what data is, where it comes from. How models are generated and how they can be adjusted. Right now, more than ever, we need to train students, but certainly also in our education for professionals, people to be continuously involved. You cannot keep standing on the sidelines,’ says Huysman.
Marleen Huysman and Marlous Agterberg are founders of the part-time programme Managing AI Wisely, which will start in January 2025. This programme to managers who have already started or are planning to develop and/or implement AI in their organisations. Check out the programme here.
VU School of Business and Economics for Professionals
With a programme from the School of Business and Economics for Professionals, you choose high-quality, innovative, and ambitious education. With a meaningful impact on the economic, social, and global challenges of our time. Always part-time, so it can be easily combined with work and personal life. From long-term programmes like MBAs and master's degrees in various disciplines to short-term courses focused on specific developments such as sustainability or data-driven management, as well as more general topics like leadership and change management.
SHARE THIS PAGE
