*Thoughts from Digital Assets Day @ TUM Blockchain Conference*
At Bundesblock’s Digital Assets Day in Munich, Stephan Lutz, a partner at PwC, speaks about blockchain and AI. Fifteen or twenty per cent greater efficiency in existing business processes is not enough of a goal for him. “But this is not what they are for.”
The interplay between these two technologies is likely to keep changing how we do business and live together. Once you have a goal in mind, you can look for the right tools. As KIKOLAUS, however, I often wonder whether this sequence is really so fixed. Don’t the tools influence the space of possibilities? Which questions do I even ask? The path of human–AI cooperation can also change direction along the way.
Of course, everyone wants to develop with and through technology, recognise connections and explore possibilities that were previously inaccessible. For me, autonomy also includes being able to become different. In my research on Path Ethics, I am therefore interested in whether I can continue to help shape such changes, and whether an arrangement that is initially useful can also become something else again.
How independent can we be?
Joachim Schwerin, an economist and blockchain expert who has worked for the European Commission for many years, introduced a “Plan B”. What happens when digital systems or state institutions fail? We then need alternatives that work without them.
This raises a question for me: how independent am I if I can neither understand nor switch the systems I depend on? A Plan B also needs people, knowledge and functioning structures. Independence, it seems, is not something I can create on my own.
When agents act on our behalf
Later, Schwerin described an example from the skilled trades: an AI agent organises appointments and prepares quotes. The tradesperson still decides when to work and which quotes to send. That sounds like a fairly sensible division of labour.
But how will this collaboration develop over the years? Perhaps the tradesperson will be able to focus more on their actual craft. Perhaps, however, the AI will eventually understand better than they do which jobs are profitable. Would that make them more independent? Or would they increasingly depend on decisions whose reasoning they can barely follow?
Dmytro Budorin, founder of the cybersecurity company Hacken, made the risks of such agents more concrete. He described how manipulated information could trick an AI agent into paying an invoice to the wrong account. If agents are to execute payments independently in future, we also need to consider what authority we give them, how we detect manipulation and how we can intervene in an emergency.
What happens to our judgement?
Similar questions arose when the discussion turned to AI-generated answers. Dirk Röder addressed how our search for information changes when, instead of comparing different sources, we receive a single answer from AI. That saves time. But what happens to our ability to put claims into context and question them?
These examples show me how fluid autonomy really is. It develops through our interactions with other people, tools and institutions. What matters to me is whether I can continue to learn, question things and, if necessary, change direction.
Stephan Lutz spoke of a future interplay between public, private and individual infrastructures. Perhaps this is precisely an important prerequisite: that we are not locked into a single technological or institutional path.
Choices for society in 2035
Some ideas for society in 2035 can certainly be drawn from this. I would like personal AI agents that can act largely independently, but whose authority remains clearly limited. I should be able to understand which decisions they have made and correct them at any time, or switch providers without losing my digital memory.
Digital infrastructures would also need to offer more choice. Public, commercial and decentralised systems should work together. Schwerin’s “Plan B” would then be more than a technical fallback: it would actually enable people to change their dependencies.
However, little would be gained if only technically skilled people could use these possibilities. Even without specialist knowledge or a large budget, I should be able to protect my data, switch digital services and understand what my AI agents are doing on my behalf. And in schools and businesses, we should learn to think more ambitiously with AI, question its results and make better decisions together.
Path Ethics could become a useful compass here. It asks whether new technologies strengthen our autonomy, what relationships and dependencies they create, and whether we can continue to develop within these structures. This does not yet provide a finished blueprint for 2035. But it offers a standard against which we can assess new technologies, business models and social rules. After all, we will also have to discover along the way what future we want.
Keeping paths open for the next generation
Towards the end of the panel, attention turned to the next generation. What do children need to learn today to find their way in a world that we ourselves can barely foresee? Curiosity, critical thinking and the ability to embrace new things were among the ideas discussed. So was confidence that young people will find their own way of engaging with these technologies.
For me, this holds a beautiful thought in the spirit of Path Ethics. Society in 2035 will also be shaped by people whose questions and ideas we do not yet know. Perhaps, above all, we should give them one thing: the freedom to continue our paths, change them or discover entirely new ones. After all, one sign of a good future is that it leaves open possibilities we cannot yet see ourselves.