Sign in Subscribe
Machine learning

Researchers at University of Zurich Explore Artificial Intelligence and Law

Strategic Insights

— 1 min read

Scientists at the University of Zurich have embarked on an innovative study to investigate the capabilities of artificial intelligence (AI) in classifying legal interpretations. Utilizing a unique dataset and Large Language Models (LLMs), the research aims to determine how AI can be employed to identify complex features relevant to the legal community. By comparing the performance of proprietary and open-source models, the study demonstrates the potential of AI in improving resource- and time-efficiency in legal analysis.

Key Takeaways:

  • Researchers at the University of Zurich have conducted a study on the application of Large Language Models (LLMs) in classifying legal interpretations, with a focus on the European Court of Human Rights (ECtHR).
  • The study utilized a unique dataset and obtained significant results, implying that feature-extraction using LLMs leads to robust outcomes and allows for greater resource- and time-efficiency compared to human annotation.
  • The research used methods such as few-shot and zero-shot chain-of-thought prompting combined with self-consistency to evaluate the performance of proprietary and open-source models.
  • The study's findings imply that Large Language Models can play a more significant role in the extraction of complex features relevant to the legal community.
  • The research has been reviewed and provides insights into the potential applications of AI in the field of law.

Statistics:

  • The study utilized a unique dataset to analyze the performance of Large Language Models in classifying legal interpretations.
  • The research found that feature-extraction using LLMs leads to robust outcomes, with a significant reduction in time and resources required compared to human annotation.
  • The study compared the performance of proprietary and open-source models using methods such as few-shot and zero-shot chain-of-thought prompting combined with self-consistency.
  • The research implies that Large Language Models can extract more complex features relevant to the legal community, with significant implications for the field of law.

Sources:

  • VerticalNews
  • University of Zurich
  • Gaspar Dugac, University of Zurich, Ctr Legal Data Sci, Pestalozzistr 24, Ch-8032 Zurich, Switzerland
  • Artificial Intelligence and Law (Springer)
  • Springer (www.springer.com)
  • Artificial Intelligence and Law (www.springerlink.com/content/0924-8463)

Read more