IND

Van Doesburg was working at the Dutch Immigration and Naturalisation Service (IND) during a major digitalisation drive. Files containing asylum applications were being updated and digitised. Automation was intended to make civil servants’ work easier, particularly in the event of legislative changes. But when the system went live in 2012, Van Doesburg realised it wasn’t quite that simple.

‘People often assume that legislative changes are straightforward: the automated software will simply implement them. The problem is that when you discuss legislative changes with lawyers, IT specialists and policy analysts, within five minutes they’re shouting at each other. They all think the law means something slightly different, and they all use their own jargon.’ Van Doesburg noticed that both the Tax Administration and the IND faced frequent legislative changes. In his PhD thesis, he set out to find a solution that could work for both organisations.

Did you manage to find a solution?

‘Yes. In my thesis, I propose a model to make it easier to reach agreement on the meaning of the law. The key question is: what do you actually use legislation for? You use it to steer people’s behaviour. People are constantly acting in accordance with the rules. The discussion about a legislative amendment becomes much simpler if you discuss the text in those terms. It is also a way of speaking that aligns well with how machines work: this is the current situation, and under these circumstances, something different will happen from now on.’

‘Developing the model was actually the easiest part of my PhD research. The challenge was convincing lawyers, IT specialists and other experts of the value of having a shared model for interpreting legislation and regulations.’

How do different experts view legislative amendments?

‘A lawyer would say: “In Article so-and-so, that section is removed and a new paragraph is added, which reads as follows …” If an IT specialist reads that, it drives them mad. I’ve never heard of anyone who has to implement legislation, who turns up at work and, upon reading a legislative amendment, knows exactly what needs to be done. But if you are implementing legislation, you do need to know that. How much tax does someone have to pay? Will someone be granted a residence permit?’

How is it that it isn’t clear to everyone exactly what legislative changes entail?

‘That surprised me too. The legal system functioned very well without computers. So why does digitalisation make things more complicated, rather than simpler? Digitalisation places entirely different demands on a legislative amendment: the meaning of an amendment must be translated so a machine can understand, while at the same time people still need to be able to check the outcome afterwards.’

‘The law has long functioned on the basis of implicit interpretations. As long as you’re among peers, that’s not a problem. It only becomes complicated when people from other disciplines get involved.’

But aren’t lawyers known for being very precise in how they put things in writing?

‘Lawyers are precise within the context of their peers. However, it is no longer simply a matter of explaining your interpretation to another lawyer, but to other stakeholders as well – someone who develops a computer system, for example. In practice, it is not the lawyer who says: “This is how we work, and this is what I need.” Instead, software engineers say: “These are the methods we use; how does your work fit into them?” IT professionals tend to view digitalisation as a one-size-fits-all model: what works in one discipline will probably work here too. Lawyers, generally speaking, remain rather passive when it comes to digitalisation. They expect IT specialists to solve the problem for them. And when that doesn’t work, they complain about it. In my thesis, I want to show that it is important to explain the essence of the law. I think that helps to explain to others what requirements a digital system for lawyers must meet.’

How does your research help with that?

‘In my thesis, I propose a way for all stakeholders to reach agreement on the meaning of the law. In this way, you can build a system step by step: within this computer system, these words have precisely this meaning. If someone then comes up with a different interpretation, they’d need to make a very strong case. The more people who support a particular interpretation, the stronger it is. The more disputes there are over an interpretation, the weaker that interpretation is. That is the potential of this type of model.’