“AI appears to make justice more accessible, but that promise is often not kept,” researchers say

They are philosophers first, lawyers second. Cees Zweistra, university lecturer in law, ethics and technology, warns the reporter. Before the interview has even started, he has already started talking about his favorite philosopher: Martin Heidegger.

“Technology is a way of looking at the world. A compelling frame in which everything appears as something that should be useful” – Heidegger already knew that in the 1950s. In his office at Rotterdam Erasmus University, Julie Hoppenbrouwers, PhD student on disruptive technologies in legal practice, is still waiting. They wrote the book together Faster & Better? The issue of technology and AI in legal practice.

Preliminary, unpublished research from 2025 from the Center for Law, AI & Design shows that more than 90 percent of lawyers surveyed use generative AI systems, such as Copilot, Harvey and CoCounsel, for structuring ideas and writing summaries, among other things.

Why Heidegger?

Zweistra: “Martin Heidegger wrote in his The question of technology (1954) that a hydroelectric power station turns the Rhine into something that has no value in itself, but is only a raw material.”

That fate, becoming a raw material, will also affect lawyers who use AI, Zweistra thinks: “Just like with the hydroelectric power plant, AI is not a neutral instrument that we use for a purpose. Technology, and AI in particular, changes our view of the world, causing both nature and people to become increasingly instrumentalized. Translated into legal practice: lawyers become consumers of systems instead of independent professionals.”

At the same time, AI does not necessarily make people more efficient, you write. How about that?

Hoppenbrouwers: “You see it everywhere: in the judiciary, at conferences, symposia and LinkedIn: the frame of the ‘innovator’s dilemma’. The idea is that as an organization you have to be good at using new technologies. If you don’t do that, you will be overtaken by others. A kind of business Darwinism.

“Unjustified; in practice it appears that efficiency improvements are not at all self-evident, because behavior adapts. Take the difference between a word processor such as Microsoft Word and a typewriter. On a typewriter you have to start over if you make a mistake, while in Word you can make immediate adjustments. That seems more efficient, but you see that people spend more time endlessly adjusting and rewriting texts, which ultimately takes them longer.”

Zweistra: “We argue that technology is not a simple instrument, but works paradoxically. The means with which you think you can solve a problem can have an attractive effect. In our Master’s degree in Law & Technology, students developed an AI tool for the municipality of Rotterdam for the accelerated handling of WOZ objections. The result: a sharp increase in objections, partly because AI lowers the threshold for submitting them. The promise is that AI will help to reduce workload and remove backlogs, but in practice it turns out that the opposite: new backlogs arise.”

Isn’t it positive that AI lowers the threshold and more people have access to justice?

Hoppenbrouwers: “AI seems to make the law more accessible: people can draw up objections and procedural documents more quickly. But that promise is often not kept. In practice, these documents are less good than they seem, and sources have to be checked more extensively, which creates extra work for both the other party and the judge.

“The judge has now indicated that this can even amount to abuse of procedural law, because with the help of a tool you place extra work on others. The question is therefore whether AI is the right tool to improve access to justice.”

Aren’t these mainly usability problems, and can AI be useful when used correctly, for example for routine tasks?

Hoppenbrouwers: “The idea that problems with AI are mainly caused by incorrect use, or will disappear by themselves as the technology improves, is exactly the view we criticize. You saw the same thing with the rise of social media: we thought that problems such as addiction would disappear if we learned to deal with them better. That did not happen.

In the Netherlands, only about 6 percent of statements are published, much other relevant material is protected by copyright, Zweistra says. “Many legal tech entrepreneurs argue for publishing more so that they can use it in their models. But we argue against this: it is not the extent of case law, but the selection that is important. That is what it is about.”

Zoom in

You write that AI can put pressure on the independence of lawyers. How does that happen?

Zweistra: “You have seen for some time in the judiciary that dependence on technology can be a problem. Take the KEI project, a large digitization program that caused the judiciary to get into financial problems and ultimately failed in 2018. One of the causes was that supplier Oracle could no longer or did not want to supply a crucial part. This brought the entire project to a standstill.”

“That shows how vulnerable you are if you do not have your own IT infrastructure. This demand for strategic autonomy has been around for some time within the judiciary. You now see that more attention is being paid to this: the judiciary is working on its own language model and its AI strategy states that the judiciary itself must maintain control over both the technology and the underlying infrastructure.”

“At the same time, this applies much less to legal work in the business community. There, commercial language models are widely used, often from large foreign parties. It is precisely there that the dependence on the models is perhaps even more important.”

The authors in the book do not provide a recipe for how to deal with AI. They do state: AI is a selectively used tool, and legal practice does not actually need it. Zweistra even wants to go so far as to suggest that lawyers take the meaning out of their profession by using too much: “If you look at the ‘material’ with which lawyers work, the cement and stones of their profession, they are texts. By collecting texts and reading them in specific contexts, lawyers create new meaning.”

Give a gift





Part





Email the editor