Prof. Dr. Frens Kroeger is an independent consultant in the field of organizational development and technology. He has worked at universities in the UK, Switzerland, the USA and Japan, led research projects on four continents and now advises companies ranging from SMEs to multinational corporations. In the ARIC interview, we spoke to him about trust, AI and ecosystems for innovation.
ARIC Hamburg: Who are you and what do you do?
My background is in trust research, with a focus on organizations and technology. This is actually quite a rare combination, because trust has long been thought of primarily between people and less in connection with technology.
Today I work as an independent consultant and deal with precisely this interface: How do organizations deal with new technologies, and when are these systems actually truly trustworthy?
What does your work consist of?
A lot of my work comes from organizational development and information and cyber security, and now more and more from dealing with AI.
In practice, I often talk to companies who say: We know that AI is important, but what does that mean for us in concrete terms?
So: What can AI really do for us? Which systems suit us and how do we integrate them in such a way that they are not just an experiment but bring real added value? What data can and should we use at all? And are we as an organization even ready for it, i.e. are we “AI ready”?
And this often brings us very quickly to questions about Trustworthy AI.
What was your path to Trustworthy AI?
I started working with trust early on – since my PhD at Cambridge – and then later increasingly worked with tech companies. I saw first-hand how much AI can change everything, at companies like IBM, Google, DeepMind and OpenAI. And trust was always a huge issue.
It quickly became clear to me that the real question is not just what these technologies can do, but how organizations deal with them and whether they do so in a way that works in the end.
You write on your website: Trust is at the root of everything we do. It is crucial to successful business, policy and technology. Why do you think trust is so important?
Surprisingly little works without trust, especially in complex situations.
And that is exactly what our world is like: we are constantly dealing with things that we cannot fully understand or control. If we wanted to secure everything – with knowledge, with rules, with contracts – we would never be able to take action. Trust is basically what allows us to make decisions anyway.
And that is why it is so central, in companies, in politics and especially in technology. This is particularly visible with AI, because it brings a new level of complexity and has a very specific influence on the decisions we make.
What exactly does trust actually mean?
Essentially, trust means that we deal with uncertainty, i.e. that we act even though we cannot rule out all risks.
But very important: it is not simply a “feeling”, it must not be blind trust. It must always be based on certain foundations: on information, on experience, on well-founded expectations about how someone or a system will behave.
And that is precisely why trust can be justified – or not.
What can we take away from this for current AI debates?
For the AI debate, this means above all that neither excessive trust nor blanket mistrust are helpful.
We need to put our trust on a firm footing and actively lay the foundations for it. In other words, understand: Can a system really do what we expect of it? Does it handle our data responsibly? And does it work efficiently – or does it produce results that initially seem plausible but later turn out to be incorrect? Because all of this determines whether trust is justified or not.
And that is precisely the idea behind Trustworthy AI.
What role do ecosystems play when it comes to trustworthy AI?
A very big one. Ecosystems are important because they create the basis for ensuring trustworthiness in the first place. After all, AI use does not happen in a vacuum, but in the interaction of many players: providers, platforms, data sources, internal and external partners. And trustworthiness must run through this entire chain.
The only problem is that this complexity also creates an enormous lack of clarity. The crucial question is often not whether there are solutions, but which ones can be used sensibly – and how you can tell whether they are really trustworthy.
And this is precisely where I believe there is a key gap, which I am also working on with a partner. Namely, creating orientation for the very practical question: Which forms of AI are really trustworthy in a specific corporate context?
What can an interdisciplinary center like the ARIC do?
An interdisciplinary center like ARIC can make a real difference at precisely this point. We all need to work together to get to the bottom of such complex issues, and ARIC is particularly strong at bringing the right people together.
This is particularly crucial for a topic such as Trustworthy AI, because it is not just a technical issue. It is also always about organization, processes and decisions.
And ARIC can help to develop orientation – not only to share knowledge, but also to create a common understanding of what works in practice and what does not.
Anything else you want to add?
Above all, I would advise companies: Don’t wait too long, and don’t just get stuck with the LLMs, i.e. ChatGPT and co. Companies that are now getting serious about AI are building up real advantages right now.
At the same time, however, you shouldn’t rush in blindly. With AI in particular, it is often the selection and type of integration that determines whether it will really work in the end or whether you will end up with problems.
And that is precisely why Trustworthy AI is not a “nice to have”, but the basis for AI to really work in everyday life.
Interview: Sabrina Pohlmann
With our interviews, we want to introduce you to different perspectives and players in the field of AI. The positions of our interview partners do not necessarily reflect the positions of the ARIC.
Further interviews:
