Introduction to AI Ethics

Language of instruction

English

Course content

The course introduces the fast evolving interdisciplinary research area of Artificial Intelligence (AI) Ethics to doctoral students who are interested in either AI as a computer science discipline or students interested in researching the societal and personal impact of AI technologies introduced in society.

The course consists of 7 modules.

The first two modules establish a common foundation for following the course.

The first introductory module presents AI as a technical discipline, the main methodology and research questions. The second module offers an introduction to artificial intelligence studies in social sciences and organisational studies.

In module 3 the students learn the foundations and state-of-the-art in accountability and transparency of AI. They get an entry point to research in this area: learn how to learn more and how to engage with that research community.

In module 4 the students will learn what the basic problems and approaches are to making algorithms explainable. They will understand the difference between explainability and interpretability of algorithms and how transparency relates to these concepts. The students learn to compare and evaluate different ML algorithms with respect to their explainability.

In module 5 the students will learn what the concept of fairness means with respect to algorithms. They will learn to recognise the different definitions of fairness, their motivation, strengths and weaknesses. the students will be introduced to the basic methods for mitigating bias in algorithms and data (preprocessing, in-processing and post processing).

In module 6 the students learn the role that privacy concerns play in artificial intelligence. In particular they will be introduced to the basic principles and methods of ensuring differential privacy and data.

Module 7 is reserved for discussing open research problems in AI ethics, challenges and possible approaches.

Each module consists of 4x45 minutes. Lectures are combined with hands-on exercises for the students and discussions. Reading assignments will be assigned between each module.

Learning outcomes

For non-computer science students the goal of the course is to enable students to identify the way in which algorithms malfunction: do they malfunction for individuals, groups, in a given context, on a given task. This helps algorithm designers address the problem.

For computer science students the goal of the course is to understand the implications of AI for society and organizations, learn the current debates about the design and use of AI, develop a critical understanding on AI.

For both groups the learning outcomes are:

Knowledge.

• Identify the basic problems studied in explainability, fairness, accountability and privacy in AI.
• Interpret, explain and extend the need for, and challenges of, AI Ethics accoross social sciences and computer science.

Skills.

• Appraise the ethical aspects of AI problems.
• Match a specific AI Ethics challenge to its most relevant discipline.
• Ability to find AI ethics literature across disciplines

General competence.

• Reading and explaining scientific articles.
• Research project management.
• Scientific reporting.

Study period

Autumn 2021

Due to the pandemics the course is planned as an online event. If conditions improve a live event will be planned. The course is planned as a two week event.

Credits (ECTS)

10 ECTS

Course registration and deadlines

31 August 2021

Register here .

Recommended previous knowledge

None

Compulsory Requirements

Participation in the assignments and discussions in the modules is compulsory.

Form of assessment

The students will be assigned a 2 month project, individual or in pairs. The project proposal will be developed together with the lecturers (or guest lecturers). The students will submit a written 15 page (5000-8000 words) report for evaluation. Grades are pass/fail.

Who may participate

Doctoral students from norwegian universities

Programme

Contact

Department of Information Science and Media Studies

Academic responsibility

Marija Slavkovik

marija.slavkovik@uib.no

Lecturers

The course is given jointly by Marija Slavkovik (UiB) and Miria Grisot from the University of Oslo. The students can register either at UiB or at UiO. The lectures will include invited guest lecturers from among the international academic researchers in AI ethics, as well as NAV and Datatilsynet.

Reading list

Course location

Virtual