Course // Responsible AI, Law & Society

Spring 2020

Overview. The deployment of Artificial Intelligence systems in multiple domains of society raises fundamental challenges and concerns, such as accountability, liability, fairness, transparency and privacy. The dynamic nature of AI systems requires a new set of skills informed by ethics, law, and policy to be applied throughout the life cycle of such systems: design, development and deployment. It also involves ongoing collaboration among data scientists, computer scientists, lawyers and ethicists. Tackling these challenges calls for an interdisciplinary approach: deconstructing these issues by discipline and reconstructing with an integrated mindset, principles and practices between Data Science and Law. This course aims to do so by bringing together senior undergrad and grad students from both disciplines into teams that work together on joint tasks in a series of four-hour long workshop-style meetings.



Prof. Avigdor Gal
Faculty of Industrial Engineering & Management

Prof. Niva Elkin-Koren
Faculty of Law
University of Haifa

Teaching Fellows

Guy Berkenstadt
Data Science Master student

Karni Chagal-Feferkorn
Law PhD student
University of Haifa

Shlomi Hod
CS PhD student
Boston University

Audience. Multidisciplinary: approximately half the students are pursuing an LLM (Masters degree) in Law and Technology from University of Haifa.The remaining half are Data Science & Engineering senior undergrad students from the Technion.

Structure. The course comprises six meetings of four hours, in a workshop format. The topics explore some of the core issues in the landscape of AI, law and society.

Class Topics Verticals
1 Liability & Robustness Autonomous Vehicles
2 Discrimination & Fairness Labour Market
3 Transparency & Explainability Finance (credit score)
4 Privacy & Surveillance Geolocation
5 Integration: Content Moderation Social Media Platforms
6 Project Presentations and Summary

Learning Objectives. At the end of the workshop, the students will…

  1. … be aware of the impact of AI on individuals, groups, society and humanity, and proactively spot ethical issues and scan for unintended consequences and potential harms.
  2. … be able to communicate with professionals from other disciplines, identify gaps in the meaning of terms and perspectives, and develop a shared language.
  3. … possess introductory knowledge and skills to oversight and audit AI systems through their life cycle (design, development and deployment).
  4. … be able to find and use resources to achieve all of the above.
  5. … take the first steps in shaping their responsibility as professionals, and be motivated to act upon it.

Format. The teaching is based on the signature pedagogy of each discipline; case-studies for Law and iterated and interactive research of data (e.g., with Jupyter Notebook) for Data Science. These two pedagogies are being used in every class, accessible to all of the students, and integrated together.

Tasks, Project and Evaluation. Every class is built around one central task that requires integration of law and data science perspectives. The tasks are performed in teams which will be formed before the start of the course. Each team consists of students from both disciplines.
Teams will present their work at least twice during the course, and all of the teams are required to submit a memo/presentation and a Jupyter Notebook at the end of each class. In their final project, the teams will be asked to develop a new case-study which makes use of data sets and data science techniques to demonstrate a legal\ethical dilemma regarding responsible AI, Law and Society.
The evaluation criteria will be focused on the integration of the law and data science perspectives.