Introduction
In a fast-changing world where technology and innovation are advancing at a rapid pace, no industry is left untouched. While many sectors have already reaped the benefits of digitalization and technological advancements, the legal sector is struggling to keep up. But there are glimmers of hope on the horizon, and the Liquid Legal Institute e.V. (LLI) is at the forefront of this change.
The think tank represents an open and interdisciplinary platform that aims to promote new ways of thinking in the legal field. Founded by seven experts from the fields of legal and business practice, design thinking and computer science, it strives to bridge the gap between legal practice and technological progress.
To explore the exciting opportunities, challenges, and impacts of generative AI and large language models (LLMs) in the legal field, LLI launched a virtual roundtable in early 2023 – the “Roundtable Generative AI (ChatGPT)”. Under the direction of the authors of this article, experts and interested parties have been meeting every two weeks since then, both to shed light on the technology behind these models, and also to discuss their possible applications and ethical concerns in everyday legal life. An average of 40 participants received input from an expert at each of the appointments. Afterwards, there were intensive discussions and exchanges and ideas.
The insights gained and discussions held over the course of these roundtables were not only of academic interest. They also offer concrete application aids and food for thought for professionals who aim to take up the challenge of revolutionizing legal practice and leading it into the 21st century. This article provides an overview of the key topics and takeaways from these inspiring roundtables.
The topics of the different sessions
Dr. Bernhard Waltl and Dr. Dierk Schindler, LLI co-founders and (legal) experts in the field of AI technology, started the series and discussed the technological foundations and the potential of such models for lawyers. The spectrum of their applications ranges from the analysis of documents to contract review in legal departments and law firms. However, they also emphasized the legal risks associated with the use of LLMs, such as possible biases and errors in the data used for model training. The need for collaboration was highlighted to better understand these models and unlock their full potential.
Sebastian Nagl discussed experiences in the application of language models from the point of view of large law firms, as well as from his own research and teaching at the Technical University of Munich. He gave an outlook on further development and the possibilities that will arise from the application of larger models.
Prof. Dr. Daniel Braun gave an overview of the functionality of LLMs and their fine-tuning for specific tasks. He shed light on how these models, such as the GPT family, can be optimized for such tasks. This presentation provided insights into the latest methods of fine-tuning and their importance in practice. He also discussed the possibilities and limitations of recognizing texts created with generative AI.
Henrik Wehrs and Dr. Jens Wagner from the law firm Allen & Overy described the possibilities and limits of the use of generative AI in large law firms. They described use cases for which the technology can already be applied, such as translation, classification of texts, summaries and the automatic answering of questions.
Prof. Dr. Patrick Glauner discussed the challenges of any possible regulation of generative AI, especially with regard to the EU AI Act. This regulation was proposed by the EU Commission to regulate the use of AI technologies. Prof. Glauner not only discussed the risk of overly far-reaching regulation, but also explained his approach of not prohibiting students from using ChatGPT, but rather requiring it as part of the task.
In his presentation, Philip Hanke from the Swiss publisher Weblaw AG emphasized the importance of LLMs and AI for legal publishing companies. He explained how this technology can transform the legal publishing market by enabling more accurate and efficient processes for content creation, editing, and distribution. Philip showed how LLMs and AI can automate tasks, reduce errors and improve the quality of legal content.
Tom Braegelmann from the law firm Annerton dedicated himself to the exciting field of “Legal Prompt Engineering” and illustrated the importance of the use of LLMs and AI tools in the legal industry. His presentation showed how browser plugins that integrate the functionality of ChatGPT and LLMs can increase efficiency, accuracy, and productivity in contract reviews.
The possibilities of Microsoft Azure in the area of AI services were highlighted by Bernhard Waltl. This presentation illustrated how developers can make these services accessible via Python packages, for example, to create chatbots. The presentation highlighted the benefits of cloud-based services such as Azure, including scalability, security, and ease of use.
Dr. Sven J. Körner from the company ThingsThinking offered some perspectives on the hype surrounding LLMs and explained the functionality of neural networks, the transformer architecture, other (future) models and the human brain. He gave an outlook on possible applications and also gave the participants the practical opportunity to train an LLM themselves – which runs on their own smartphone.
Isabelle Biallaß, head of the Legal Tech & AI think tank in the judiciary of North Rhine-Westphalia and judge at the local court in Essen, led a discussion on the legal and ethical limits of the use of LLMs in the judiciary. She emphasized the increasing role of generative AI models in the legal field and discussed their opportunities and risks, taking into account the German legal situation. In particular, the aspect of out-of-court dispute resolution through such systems, as well as opportunities for access to justice in developing countries, were discussed.
Jonathan Laidler from IP Quants bid farewell to the roundtable for the summer break with his presentation. He outlined the work of the future for legal and IP professionals by automating and standardizing workflows through digital solutions, highlighting how Generative AI is already transforming the work of IP professionals. Examples illustrated how Generative AI can automate document reviews and key insights.
Takeaways
he LLI Roundtable has created a forum to discuss generative AI and its practical applications in the legal field. The presentations and discussions have produced diverse perspectives and innovative ideas that prepare the participants for future challenges and opportunities. This forum highlights the need for collaboration between technology experts, lawyers and academics to realize the full potential of generative AI while creating ethical and legal frameworks. The future of the legal industry will benefit from this advanced technology, and the LLI Roundtable can help guide this transformative development legally, ethically, and technologically. After the summer break, the roundtable will continue with new speakers and topics to further discuss current developments in generative AI and its application in the legal field. Participation is open to all members of the LLI.
Editor’s note:
This article was also created with the active support of ChatGPT, especially the summaries of the individual sessions, and the outline offered good preparatory work and enabled faster writing.