The NRP must do more than just describe the problems; it must deliver solutions

From the perspective of the legislature, what is the main problem with digitalisation?

The progress of digitalisation is opening up many applications for the administration – from a purely technical point of view. But is there also a legal foundation for these applications? This principle of legality must be adhered to, and at the same time people want to be flexible, creative and innovative. This is an area of tension.

Even on the legislature’s side, not all authorities agree on all points as to how tightly or loosely laws should be formulated to allow them to incorporate future as well as current technical developments.

That’s the legislature’s point of view. Where do you see the challenge with regard to the general public?

Here, everything revolves around trust. That’s pivotal. We saw this recently in the vote on electronic ID cards. The public was wary, and the bill was rejected.

Could it be that there is a gap between the expectations of the authorities and those of the public?

I wouldn’t put it that strongly. But the authorities have learned. The revised version of the data protection legislation has just been approved. The aim of this is to strengthen trust by better protecting data. But here, too, there are different attitudes even within the federal administration. How much data protection is necessary, and how do we nevertheless remain flexible with respect to new technical developments? The Federal Council needs to balance this out.

Let’s talk about the digitalisation of legislation and jurisprudence. Will artificial intelligence replace lawyers in future?

That’s more along the lines of science fiction. But AI could certainly take some of the strain. For example in terms of help for victims, especially when it comes to questions of compensation. Here there are many tables that list what compensation has been awarded to a victim in which cases and under which circumstances. Today, people have to trawl through countless cases to determine a compensation sum. If the data were appropriately organised, an algorithm could handle this work. Even tax returns could be read and appraised by machines.

Is it not tricky if computers take on legal responsibility?

They won’t be taking on responsibility, of course. AI will only make a pre-selection and provide recommendations. Decisions will always be made by experts. The legal specialist must weigh up whether a special situation is present in certain cases. And if a software package, for example, issues a tax decision, that must be transparently declared so that the taxpayer can call for a review by a human being.

So you are not expecting digitalisation to bring any quantum leaps in the judicial system and legislation?

No; it will take routine tasks off our hands and thus be a tool that helps us do our work more expeditiously – primarily in straightforward cases.

Whether that will then create new problems and unequal treatment, I don’t know.

What do you expect from NRP 77?

People always talk about AI in very general terms, and too seldom of specific applications. So we don’t need still more generic statements on AI and its risks.

There are several projects exploring questions relating to the legal framework, trust in cyber security or the role of AI in court decisions. For this research to benefit us in legislative practice, the results cannot be limited to just describing the problems. We know that artificial intelligence can lead to discrimination as well. That is widely known. I want to know how that can be prevented

NRP 77 projects related to the topic