Back to overview

(Ro)Bot-Human Interaction - A Digital Shift in the Administration of Criminal Justice? Substantive Law, Procedure, & Verdicts in Ambient Intelligent Environments

English title (Ro)Bot-Human Interaction - A Digital Shift in the Administration of Criminal Justice? Substantive Law, Procedure, & Verdicts in Ambient Intelligent Environments
Applicant Gless Sabine
Number 184895
Funding scheme Project funding
Research institution Juristische Fakultät Universität Basel
Institution of higher education University of Basel - BS
Main discipline Legal sciences
Start/End 01.01.2020 - 31.12.2023
Approved amount 576'499.00
Show all

Keywords (11)

Artificial Intelligence; robot-human-cooperation; legal narratives and criminal verdicts; comparative criminal law; robotics and criminal law ; digital evidence and defense rights; automated driving and criminal liability; mens rea; driving automation; human robot interface ; criminal liability and negligence

Lay Summary (German)

Könnte ein Roboter Mitschuld an einem tödlichen Verkehrsunfall tragen, wenn er den menschlichen Fahrer nicht rechtzeitig warnt? Erhöht sich die Schuld des Menschen, wenn er eine Roboterwarnung nicht beachtet und übersteuert? Dürfte der Roboter dann als Zeuge in einem nachfolgenden Strafverfahren auftreten? Wie begründen Gerichte Verurteilung oder Freispruch, wenn nicht ein menschlicher Fahrer, sondern digitale Fahrassistenten ein Unfallauto gesteuert haben?
Lay summary
Diese Fragen sind Gegenstand des Forschungsprojektes, das am Beispiel der Fahrautomatisierung insbesondere drei Problemstellungen untersucht: (1) mögliche Modifikationen der strafrechtlichen Zurechnung, insbesondere im Bereich der Fahrlässigkeitsdelikte, wenn Menschen die Warnung von digitalen Fahrassistenten (z.B. Müdigkeitsassistenten) übersteuern; (2) Reformbedarf im Strafverfahren, wenn sich Menschen künftig gegen «Maschinenzeugen» verteidigen müssen, weil ihnen die durch Fahrzeugassistenten gespeicherten und ausgewerteten Daten als Schuldbeweise entgegen gehalten werden; (3) neue Narrative in Urteilsbegründungen in Fällen, in denen Roboter im Grunde genommen als «Kopiloten» auch vor Gericht stehen müssten, aber nur Menschen angeklagt werden (können).
Direct link to Lay Summary Last update: 01.11.2019

Responsible applicant and co-applicants


Associated projects

Number Title Start Funding scheme
167182 Legal Challenges in Big Data. Allocating benefits. Averting risks. 01.02.2017 NRP 75 Big Data


Criminal law was designed by humans to create order amongst ourselves. Accordingly, its fundamental concepts, such as the notion of free will or the concept of a free appraisal of evidence or the claim of dignity and autonomy of parties to the criminal proceeding, are uniquely human. However, this human-centric approach is at risk of losing its place as a primary guiding principle as (ro)bot-human interactions increase. Until now, such collaboration has not had special standing in criminal law despite the fact that robots in various forms, software bots as well as standalone machines, already gather and process information, draw conclusions on behalf of humans and cooperate with them, or even act autonomously as their agents. Ongoing digitalization, or the increase in ambient intelligent environments, may gradually require a change in the traditional method of establishing the elements of alleged criminal conduct and in the way in which such elements are proven in court when human-robot collaboration turns detrimental. This, in turn, may have repercussions upon the narratives in criminal verdicts. This project assesses the consequences of (ro)bot-human interactions in ambient intelligent environments. It primarily uses the example of technology surveying human drivers in modern cars in the course of driving automation. This includes software bots as well as physically independent smart devices that are made for enhancing safety in road traffic in the context of Swiss penal law. The goal is to understand legal changes and predict further impacts upon criminal law. The project will cover three working packages; packages one and two will be dissertation projects and the third will be a joint project among a group of scholars in comparative criminal law, engineering and law and legal narratives. The working packages will build upon each other, addressing the following questions:(1) Substantive Criminal Law: What are the potential changes in the assessment of subjective (mens rea) elements of crimes in penal traffic law cases, particularly when robots issue warnings to humans based on their assessment of drivers’ behavior? (2) Criminal Procedure: What are the challenges for fact-finding and evidence evaluation in criminal proceedings after fatal traffic accidents involving evidence from monitoring technology? Specifically, must defense rights be modified as a consequence to the use of certain machine-generated evidence so as to provide sufficient protection to the human driver standing trial? (3) Digital shifts and its impact on narratives in criminal verdicts: How do criminal courts assess and prove mens rea in traffic incidents? Does the approach change with driving automation? Where do different criminal justice systems, with diverging legal traditions, technology adoption and traffic regulations, differ in explaining and establishing responsibility? Are there any commonalities?Smart safety devices in modern cars provide poignant examples of everyday human-robot collaboration that foreshadows a number of specific effects upon penal law. Within the domain of substantive criminal law, the demarcation of a negligent act from an intentional crime could change entirely when the mens rea could be inferred from a person’s response to (ro)bot advice. For instance, if a drowsiness detection system alerts a driver to take a break but the driver continues, eventually causing an accident, courts may be inclined to infer negligence or even intent as a result of the driver disregarding the advice. Yet, criminal procedure might not grant an adequate defense to human drivers. In criminal proceedings the importance of human testimony, most notably a defendant’s admission, will lose standing, while machine-based evidence will gain traction. For example, if a drowsiness detection system assesses a driver as sleepy, how can the driver challenge this machine-evidence if it is presented against him in court? Finally, verdicts will have to explain responsibility sharing in (ro)bot-human interactions and how machine-generated data was evaluated against human statements to justify an acquittal or conviction. With the ongoing automation of driving, robots monitoring humans will take over a crucial safety functions, like in Take Over Requests (TOR) when a car wants to hand over after having performed the dynamic driving tasks itself and must ensure the human driver is in fact capable of taking the steering wheel. As robots will likely become more embedded in our lives, similar issues will arise in other aspects of human life.