Interdisciplinary expert pool for NLU

Challenges

The development of large language models prompts more challenges than purely technological ones. Questions of data, representation, fairness, equality, and ethics are implicit and relevant to all projects in AI Sweden’s language technology portfolio and within the NLU/NLP ecosystem more broadly. 

This project addresses the lack of a suitable response. So far, resources to tackle these issues in depth have not been readily available within the scope of technological projects. Secondly, the project is a reaction to international (mainly US-centered) debates about the risks and limitations of language models.

Project and expected outcomes

The project aims to leverage the existing knowledge of researchers from the humanities and social sciences as well as civil society actors when needed to solve concrete questions and challenges that arise in the development of foundation models.

AI Sweden will invite subject matter experts to be involved in the development pipeline of foundation models, considering questions regarding the training data and, consequently, the safe deployment as well as the social, cultural, and economic implications of foundation models such as GPT-SW3.

The project’s key innovation is to systematically and at scale involve subject matter experts in the development pipeline of foundation models, leverage existing domain knowledge, and together work on applied challenges and “big” questions based on concrete use cases and needs. The expert pool will provide a platform for knowledge exchange and readily available resources to quickly and effectively tackle data, representativeness, equality, and ethics issues.

The members of the Interdisciplinary Expert Pool for NLU

Researchers and individual members

Civil society organization representatives

Descriptions of organizations and biographies of the members of the Interdisciplinary Expert Pool can be downloaded below.

Resources

Training Material for the Interdisciplinary Expert Pool for NLU

This training material provides introductory resources for anyone who would like to gain knowledge of Large Language Models, with special attention to the ethical aspects of this technology. The material is aimed at both a technical and a non-technical audience. It consists of both essential and more advanced resources in the form of readings, podcasts, open courses, videos, and book.

The material is available on our online community MyAI. Create a free account to gain access to the material.

Alternatively, the material is available as a presentation in the link below 'To the presentation'.

This list has been compiled for the Interdisciplinary Expert Pool for NLU project.

Download the presentation:

Contact