Season 1: Digitalisation, AI & Criminal Law

In the podcast «Fokus: Strafrecht» we discuss current challenges and their significance for Swiss criminal law with experts from the legal profession and other disciplines.

Season 1 is dedicated to questions surrounding digitalisation and AI: How is artificial intelligence changing our legal system? Will it soon be passing court judgements? How reliable is it in assessing the risk posed by criminals? What do tomorrow's lawyers and solicitors need to know about AI? What is identity theft? And much more...

The interplay between neurotechnologies and AI is driving an unprecedented wave of progress. In the medical field in particular, intelligent neuroprostheses and implants offer enormous opportunities for patients. However, intensive research is also being carried out into military applications within the context of ‘cognitive warfare’, as well as into commercial uses by the major players in the big data economy. 
According toProf. Dr. Marcello Ienca , this makes a thorough examination of the ethical, political and legal frameworks all the more important. Marcello Ienca is Professor of Ethics of AI and Neuroscience at the Technical University of Munich. 
We speak to him about brain-computer interfaces (BCIs), neuro-rights and cognitive integrity, the governance model for brain data that he helped to develop, and possible criminal law implications.

“Companies that use or develop AI need clear, organisation-wide AI governance. It is key to keeping pace with rapid technological developments and to acting responsibly and in compliance with regulations,” saysDr. iur. Anne-Sophie Morand. She is a lawyer, university lecturer and a member of the Data & AI Governance Council at Swisscom. There, she leads an interdisciplinary team responsible for ensuring that the innovative potential of AI technologies is harnessed in a legally compliant, trustworthy and responsible manner. She has also authored the white paper ‘AI Regulation and AI Governance’.
We speak to her about the opportunities and challenges involved in the use and development of AI tools in a corporate context, the legal frameworks in Switzerland and the EU, and the roles and functions relating to AI governance in modern organisations.

Automated facial recognition and number plate monitoring, the covert collection of biometric data, ‘fishing expeditions’ and the retention of data without reasonable suspicion all pose risks to freedom and privacy.  
Gregor Münch  is a solicitor and co-founder of the AI training platform recht-intelligent.ch. In this interview, he discusses the impact of AI on his work as a criminal defence lawyer. He highlights which AI-supported surveillance measures are currently permitted and which are not, and uses recent Federal Supreme Court rulings to illustrate why it is important to distinguish between preventative police measures and investigations by law enforcement agencies when using AI.

It is predicted that artificial intelligence will fundamentally transform the work of lawyers in the coming years. But in what ways can AI already be usefully applied in legal practice today? What legal and practical challenges need to be taken into account? And why should legal professionals explore the possibilities and limitations of this technology? We discuss these issues with expert Simon Brun, a solicitor, lecturer in criminal law at the ZHAW and co-founder of the platform recht-intelligent.ch, which offers training courses on the use of AI for legal professionals.

A significant amount of digital child pornography is hosted illegally in Switzerland. The Swiss offence of pornography (Art. 197 of the Swiss Criminal Code) prohibits not only genuine (i.e. real) child pornography, but also synthetic child pornography. In the age of AI image generators, this is an advantage for law enforcement agencies. However, Switzerland has no mandatory reporting requirement and also lags far behind other countries in terms of the options available for the permanent deletion of data.
The lawyers Sarah von Hoyningen-Huene and Jutta Oberlin have examined the various legal issues surrounding child pornography in depth. As well as calling on the legislature to act, they also urge parents to carefully consider which images of their children they wish to post online and on which platforms.

Prevention and victim support services are available atclickandstop.ch andkinderschutz.ch, whilst kein-taeter-werden.ch offers support for the prevention of offending.

What is the likelihood that individuals will reoffend? This question is relevant across the entire spectrum of the criminal justice system, from pre-trial detention through the decision on punishment or measures to the rehabilitation of former offenders. The potential risk of reoffending is assessed using various tools and methods.
We discuss how reliable these are, what risks they entail and where AI could be used in the future with Thierry Urwyler, Prof. Dr. iur., M.Sc. in Forensic Psychology. Urwyler is an assistant professor of criminal law and criminal procedure at the University of Zurich and, thanks to his many years’ experience as a research fellow in the prison service of the Canton of Zurich, has extensive expertise in the field of criminal forecasting.

Preventing crimes before they happen. This is both the promise and the danger of predictive policing. Predictive policing is based on data-driven assumptions about which areas are more prone to crime and which groups of people are likely to be involved in such offences more frequently. 
Predictive policing promises efficiency, lower crime rates and greater safety. At the same time, there are serious concerns regarding bias, racial profiling and the phenomenon of the self-fulfilling prophecy, as the greater the police presence, the more crimes are detected. 
In addition to traditional statistical tools, AI tools are also increasingly being used in predictive policing, bringing with them new and significant risks.  
We discuss this with Sarah Summers, Professor of Criminal Law and Criminal Procedure at the University of Zurich and Co-Director of the Digital Society Initiative. 
This episode is in English.

Factors as varied as daytime sleepiness, age, ADHD, or even intoxicants such as alcohol or cannabis can affect our fitness to drive and our driving ability to a greater or lesser extent. Research director Stefan Lakämper uses the immersive driving simulator Victor – supplemented by eye-tracking and EEG measurements – to analyse, under safe laboratory conditions, to what extent people are able to drive safely from a medical perspective and under which conditions. This highly complex research vehicle is housed at the Institute of Traffic Medicine, part of the Institute of Forensic Medicine at the University of Zurich. We spoke to him about how Victor ended up on the first floor and what research projects the driving simulator enables. 

Humans and artificial intelligence deal with legal issues in very different ways. AI can access huge amounts of relevant data extremely quickly, but generally has little contextual knowledge about how the state, institutions and legal systems work, or the philosophical principles underlying them. 
How do large language models such as ChatGPT deal with legal issues, depending on the data material and prompts? How does AI respond to moral dilemmas? And are AI judgements actually more transparent than those made by judges?  
We discuss this with Yoan Hermstrüwer, Professor of Legal Tech, Law and Economics and Public Law at the University of Zurich. He is an expert in legal data science, i.e. the application of data science methods in law.

Criminals are also taking advantage of the boom in cryptocurrencies such as Bitcoin and Ethereum. Prosecuting such crimes poses major challenges for law enforcement agencies. On the one hand, cryptocurrencies, which are mostly blockchain-based and pseudo-anonymous, are stored in a decentralised manner. This means that they are not controlled by any central institution (e.g. a bank). Furthermore, they exist exclusively in virtual form and are not tied to any specific location.
This means that both the conventional determination of the place of success and the classic principle of territoriality reach their limits here. Despite recommendations from the Financial Action Task Force (FATF) and the International Cybercrime Convention, which, among other things, facilitates international cooperation in the prosecution of cybercrime, transnational prosecution remains fraught with obstacles. 
We talk to Lea Bachmann about the crimes committed with and around cryptocurrencies and the adjustments that would be possible and necessary in criminal law, particularly in international legal assistance. Together with Sabine Gless, she has written a book chapter on the subject (Sabine Gless/Lea Bachmann, § 21 Internationales Strafrecht, in: Zellweger-Gutknecht, Corinne; Tschudi, Dominik; MacCabe, Kevin, Kryptowerte, Basel 2024). Lea Bachmann is a doctoral candidate at the University of Basel. In her SNF-funded dissertation, she is investigating the limits of criminal liability when using AI systems in companies, using the example of money laundering prevention. 

Modern vehicles cause far more than just damage to bodywork in accidents. With sensors and systems now recording every movement, the question arises as to what role digital data plays today in the analysis of traffic accidents and how it can help prevent such accidents in the future.  
Crash test dummies, on the other hand, provide important physical data on the risk of injury in collisions by simulating the movements of the human body. 
How are accidents reconstructed and how is this changing with increasing vehicle automation? What can we learn from this for the prevention of traffic accidents in order to move closer to ‘Vision Zero’? We discuss with Bettina Zahnd, road safety expert at EBP.

Anyone who disagrees with the AI ruling must justify their position: AI is already widely used in the justice system in China. In many other countries it is finding its way into the work of courts, law enforcement agencies and lawyers, whether in case allocation, research for the reasoning behind judgements, the reasoning itself, or even as a tool for sentencing. Where and how can, may and should AI be used? What risks do AI tools pose, particularly for fundamental rights and the rule of law? 
We talk toProf. Vanessa Franssen, a specialist in cybercrime at the University of Liège in Belgium, about the advantages and possible applications, as well as the limitations and risks of AI in court practice.
This episode is in English.

The socio-political debate surrounding deep fakes and identity theft has not gone unnoticed by Swiss lawmakers. Since autumn 2023, the new criminal offence of identity theft has been enshrined in the Criminal Code. Legal scholars are critical of this move due to many ambiguities: under what circumstances will I be liable to prosecution in future? If I use the name, image or voice of other people? If I do so as a joke or with fraudulent intent? And last but not least: is this offence even necessary? We discuss this withGian Ege, Assistant Professor of Criminal law and Criminal procedure at the University of Zurich, using the case of Sibel Arslan/Andreas Glarner as an example.

Can I let the autopilot take over on the motorway and write a text message? Driver assistance systems are rapidly gaining ground and revolutionising driving. In some US states, self-driving cars, known as robotaxis, are already permitted. Here in this country, this will probably take some time yet. However, increasing automation is already raising many legal questions. What is permitted and what is not? Who decides on approval? And who is liable in the event of traffic accidents – the driver, the manufacturer or the supplier? In her postdoctoral thesis, Prof. Dr. Nadine Zurkinden takes an in-depth look at the stages of vehicle automation and permissible risk. In conversation with Dr. Claude Eric Bertschinger, she provides insight into the world of assisted driving.

Our own experiment at the start of "Fokus: Strafrecht" shows that large language models such as ChatGPT sometimes hallucinate. Prof. Dr. Nadine Zurkinden and Dr. Claude Eric Bertschinger check their tracks on the internet, chat about their motivation for podcasting and give you a first glimpse into the first season.