Justitia online
Chatbots are digital assistants that are available round the clock. What makes a chatbot stand out?
First and foremost is usability. Members of the public should be able to contact law enforcement agencies very easily, even via digital channels, to raise their concerns. A chatbot takes on the role of a digital assistant. Furthermore, a chatbot must be embedded within an ecosystem of other contact options to broaden the channels of access. It must not block other contact channels. In the justice system, this is particularly important in the case of urgent matters, such as when information is provided that requires immediate action – for example, in the event of an imminent robbery or an offence of abuse.
How can members of the public currently file criminal complaints relating to hate crime?
The police authorities offer various options for this. All State Criminal Investigation Offices operate online police stations. There, members of the public can file digital criminal complaints with the police by completing an online form, which also includes the relevant information regarding the duty to tell the truth, etc. They can also report posts to various NGOs and projects, which in turn cooperate with the authorities. The judiciary itself does not yet, for the most part, have its own digital platforms. This is primarily due to the legal framework. We are dealing with a complex environment here.
What led to the idea of developing a chatbot for the public prosecutor’s office as well?
To ensure that concerns in the digital sphere reach us in a timely manner, the judiciary and law enforcement should also be accessible digitally without any disruption in the user experience. At the ZAC NRW, we are responsible for prosecuting high-profile cases of digital hate crime. Our remit includes, for example, offences in which prominent figures in democratic life, representatives of religious communities, politicians or mayors are targeted on social media with hate crimes and hateful posts. In practice, a member of the public who observes this and wishes to report it does not simply go to the local police station round the corner. If we also offer digital reporting options for this, it increases the effectiveness of criminal prosecution. We shed light on the unreported cases and receive more substantiated reports. Against this backdrop, the main aim of the development work was to break down what is technically feasible into the judicial framework and to outline a concrete way forward.
How does criminal prosecution currently work at ZAC NRW?
ZAC NRW is responsible for various areas of digital hate crime in high-profile cases.
Firstly, there is the collaborative project ‘Prosecute rather than just delete’. Here, we work together with the State Media Authority, the State Criminal Police Office, several police stations and, above all, partners from the media sector – namely the major television broadcasters, radio stations and newspapers based in North Rhine-Westphalia. To this end, we have set up a fully digital reporting channel through which our media partners can submit reports when hateful posts of a criminally relevant nature appear on their platforms. We then take these cases on board and follow them up with the State Criminal Investigation Office or local police stations. To this end, we have, for example, trained our editors so that they are better able to recognise which posts might actually be of criminal relevance. Posts may be distasteful or hard to stomach, but they are not necessarily a criminal offence. We are not a censorship authority. Collaborating with media organisations improves the quality of the reports and makes their handling more effective for our media partners.
In addition, we work with the Federal Office of Justice (BfJ) in Bonn. The BfJ is the regulatory authority responsible for imposing sanctions for breaches under the so-called Network Enforcement Act. If matters of criminal relevance come to light during these administrative offence proceedings, we are notified of them by the BfJ.
The Ministry of Justice has also entrusted us with a third area of responsibility, which concerns the aforementioned cases of high-profile hate crime on social media where the hatred is directed at public figures, such as mayors or people who have spoken out on political issues.
How exactly does the justice chatbot, which was tested in a proof of concept, work?
Users can report incidents relating to digital hate crime via the chatbot. Users interact with the chatbot using natural language and describe their observations in natural language. For example, the user does not need to say: ‘I would like to report the use of symbols of unconstitutional organisations’; instead, it is sufficient to say: ‘I would like to report a swastika’.
Using the AI component for language analysis, the chatbot identifies the nature of the offence and accesses a detailed catalogue of requirements used in law enforcement practice. The chatbot works through this catalogue in a structured manner whilst conversing with the person making the report, asking, for example, for the address of the post and for further information. The chatbot therefore requires well-structured dialogue frameworks. In the background, it then automatically retrieves information from the social network to provide as much detail as possible about the facts of the case. Once everything has been gathered, the chatbot presents the complainant with a summary of the case and sets the criminal complaint in motion.
Citizens can therefore file a report at any time. What other advantages do you see?
One of the key advantages is the seamless user experience. The user does not have to leave the digital environment to file a criminal complaint. They can contact the authorities right where the offences were committed – namely, in the digital sphere. Ideally, the chatbot should be accessible via any mobile device.
When can members of the public expect to be able to file their reports via a chatbot?
So far, we are at the proof-of-concept stage. There is still a great deal of work ahead of us. However, experience to date shows that there is considerable demand and interest. We are currently in the process of clarifying whether and how this could be implemented in practice.
On a technical level, for example, we still need to refine the design and the logic of the conversation flow in order to transform the PoC into a fully functional application. We will certainly also carry out test phases in between, for instance with a limited group of participants. I am optimistic that, with a few months’ development time, we would be in a position to deploy the chatbot, both from a technological and a functional perspective. The administrative processes involved also still need to be clarified.
Can this idea also be applied to other areas within the judiciary or other administrative sectors?
The application can be extended to many other areas. I am aware of scenarios in civil justice that deal, for example, with the concept of a digital legal applications centre. In all cases, the aim is to work out how citizens can contact the justice system via a chatbot to submit an application. Numerous projects are already in the planning stage or even being trialled.
Overall, the use of a chatbot can be applied to many areas of public administration where the aim is to facilitate digital dialogue with the public. Of course, we will always need physical offices, but there are many situations where simply providing information and ensuring rapid availability are already very important. That is why I believe that, in future, we will have a model combining both digital and analogue communication.
What makes sense for the judiciary may also come to the fore in many other areas of public administration.
Thank you very much for the interview.
