Bart Custers: 'AI can predict rulings, but beware of Kafka'

The work of judges can also be done by AI systems. These systems can formulate an objective proposal for a ruling. But such proposal should always be checked by a judge, says prof. Bart Custers.

According to legal expert and technical physicist Custers, it will take a while for AI to make its appearance in a court of law. ‘In the US, these kinds of tools are already used to predict whether it makes sense to hire a lawyer. If you made a bad purchase for $ 100 on eBay, it doesn’t make sense to sue, but AI can rule quickly and cheaply.’

Bart Custers, Professor of Law and Data Science and Head of the eLaw Department

AI systems help junior lawyers

In the Netherlands, AI systems help junior lawyer doing research work for law firms: they no longer have to look up all the relevant earlier rulings and regulations themselves. Custers: ‘The problem remains that just like Google, these systems don’t deliver the same results for everyone. In class, I sometimes ask students to Google a specific item and compare their results. They quickly discover that they all live in their own bubbles.’ 

Crushed in the wheels

Besides the technical opportunities, Custers also focuses on how AI is regulated. ‘Two neighbours, one living at number 187 and the other at 187A discover that they have different home insurance payments. Turns out the insurance company concluded based on a data bank that the inhabitants of 187A had a different risk profile. Legally, this is allowed, although morally, we may take issue with it. We have to constantly adjust the rules so people don’t get crushed in the wheels of technology. We don’t want a Kafkaesque society.’ 

'  Who knows what kind of technology we will have inside our bodies in thirty years’ time?'

On the other hand, Europe doesn’t have any large tech companies like those in Silicon Valley or China because of our stricter privacy rules. Custers: ‘The EU may not want any privacy debacles like those involving Facebook and Cambridge Analytica, but now we’re missing out on opportunities. What we need is a good middle way.’ 

We need to anticipate

AI belongs to experts in technology, ethics and law
The rules of the game often lag behind the technology. ‘It’s all going so fast; who knows what kind of technology we will have on and inside our bodies in thirty years’ time? This is something we need to anticipate, which is why I try to bring together experts in technology, ethics and law. That way we can create privacy by design: build in all kinds of guarantees beforehand, to avoid problems developing. For example, at eLaw we work with doctors and ethics experts who conduct ever more detailed analyses of patient data.’ And don’t think the ethics and legal experts just wag accusing fingers and try to slow things down. ‘On the contrary, people often find out a lot more is possible than they thought.’

Text: Rianne Lindhout
Photo: Patricia Nauta

AI Leiden


More information:
Artificial Intelligence Universiteit Leiden
eLaw - Centrum voor Recht en Digitale Technologie