KIN blog

Book Launch: Managing Artificial Intelligence Wisely

Reading Time: 3 minutes
Image: Co-authors Lauren Waardenburg (top right), Marleen Huysman (bottom left), and Marlous Agterberg (bottom right), and host, Han Gerrits (top left)

November 24, 2020, marked the Digital Symposium, “AI in Practice” and launch of the new AI management book authored by KIN Center for Digital Innovation’s Lauren Waardenburg, Marleen Huysman, and Marlous Agterberg and published by Stichting Management Studies. Titled, S.L.I.M. managen van AI in de praktijk (Managing Artificial Intelligence Wisely), the book is unique in its focus on operational and implementation processes within real organizations. The book illustrates how AI is transforming work, and highlights key implications and recommendations for managers considering embarking on or already implementing AI.

The symposium was hosted by Han Gerrits (Capgemini). It featured a keynote by co-author Marleen Huysman (KIN) followed by a reflection by Michel van Leeuwen (Ministry of Justice and Security). The session ended with a panel discussion with some of the managers interviewed for the book: Ruut Durmadaneker (KLM), Daniël Meel (ABN AMRO), Annemiek Bakker (Centraal Beheer) and Xenia Kuiper (Philadelphia), as well as science journalist, Bennie Mols.

Image: Organizations featured in panel discussion and book

In her keynote, Marleen explained that AI is differentiated from previous automation efforts because it requires large quantities of data, it is self-learning (thus black boxed), and its focus is knowledge work. Marleen shared the four key challenges for implementing AI in practice which were also reflected upon by the panelists:

1. Organizing around data becomes a challenge because so data are the central building blocks of AI systems, and requires human coordination to organize for the data. Thus, not only the ability to collect data, but also to process and restructure the organization around data becomes key.

AI is more like LEGO than Playmobil…but we can no longer say that IT technologies are neutral and without bias. Especially not when it comes to AI, you can’t just leave it to your IT department. You need to engage as a whole organization - Michel van Leeuwen, Director of AI at the Ministry of Justice and Security

2. Testing and validating concerns the question of when a system is good enough to hand tasks over to. To create AI systems that “work”, not only technically but also socially, within organizations, requires continuous effort and is necessarily combined with broad and deep knowledge of use contexts and technology.

We can only learn by putting the robots in practice and let them get their flight hours…but it’s important to take it easy…to improve and change the work processes and adoption at a human tempo, to go slower rather than faster. It’s a long long term commitment Xenia Kuiper, Program manager at Philadelphia

3. Algorithmic brokering becomes a necessary and emergent role to translate and bridge AI outcomes to users.

People need to know what AI can or can’t doBennie Mols, science journalist

Our robot team works as the broker between the patients and the system developers  Xenia Kuiper, Program manager at Philadelphia

4. Changes to work processes result from the implementation of AI systems and organizations will need to be able to learn and adapt step by step.

We eventually moved from an external supplier and placed our entire architecture and management internally to keep the expertise within the organizationAnnemiek Bakker, Product Manager at Centraal Beheer

These themes are further explained within the book with one chapter dedicated to each challenge. Finally, Marleen closed her talk by introducing the W.I.S.E. framework for managers. Wisely implementing AI in organizations requires including Work-related insights, Interdisciplinary knowledge, a Socio-technical change process, and Ethical awareness (W.I.S.E.). Using the W.I.S.E. framework, managers can ask the right questions to understand the risks and trade-offs toward implementing AI wisely in their organizations.

The session ended with a lively panel discussion that touched on topics discussed in the book, such as the benefits of developing AI systems in-house or the cross-functional value that AI solutions can provide across organizations past the initial investment that can be significant. The book features these cases from KLM, Centraal Beheer, ABN AMRO, and Philadelphia in depth, as well as providing additional cases around predictive tumor modeling, predictive policing, predictive HR analytics, and a smart powerplant. Buy your copy of the book now (in Dutch).

Want to find out more?

Author: Shauna Jin