Home Events AIDI micro-workshop on global fairness and inclusiveness in AI
Community events

AIDI micro-workshop on global fairness and inclusiveness in AI

(Photo by Gabriel Vasiliu)

As conversational AI models are overwhelmingly trained on datasets of Western origins, they often suffer from a lack of data and focus concerning other regions of the world. This can produce blindspots, bias and imbalance.

In this micro-workshop co-organized by the EBU's AI and Data Initiative (AIDI) and Deutsche Welle, we explored the issue of Western bias in AI and its repercussions on e.g. public service media global coverage.

We approached this question through expert presentations and group discussions, focusing on:

  • Imbalanced representation of the global population
  • Difficulty in accessing thorough and accurate local information about non-Western regions, for both local and Western populations
  • The socio-technical mechanisms behind this issue
  • How to address the question of global inclusiveness from an ethical perspective

This approach was based on the work of DW Akademie on media development and AI.

Following calls from actions from our experts (see video), participants discussed the following questions in rotating groups: How can PSM prepare society for the changes that generative AI bring about, especially during elections? How can AI policies that lead to accountability and agency be developed? How do we push for accountability in AI on a global scale? How to we encourage crosschecks from multiple diverse experts and create partnerships? How do we counter AI bias?

Re-occuring themes included setting up partnerships with organizations and boards across various fields, encouraging a multi-stakeholders approach to governance, recruiting diverse staffs, addressing datasets diversity and updates, trusting local grassroot NLP initiatives over corporate ones, setting up inclusive guidelines, taking into account non-Western groups' (e.g. indigenous populations) IP rights, taking part in AI ethics conversations with e.g. UNESCO, and increasing AI literacy among both professionals and audiences.

A report will be published soon that discusses these questions into further details.

Below are further interesting links readings related to this topic:

Taking A Continental Leap Towards A Technologically Empowered Africa At The AUDA-NEPAD AI Dialogue

RSF and 16 partners unveil Paris Charter on AI and Journalism

Paris Charter on AI and Journalism

Ghana NLP

Lepala AI

Lesan AI

Masakhane

Speakers

Ruth Kühn

Co-organizer

Senior Technology Manager, DW (Germany)

Read biography

Jan Lublinski

Co-organizer

Head of the Policy and Learning Department, DW Akademie (Germany)

Read biography

Heather Dannyelle Thompson

Manager of Digital Democracy, Democracy Reporting International (Germany)

Read biography

Nele Lenze

Senior Lecturer, DW Akademie (Germany)

Read biography

Julius Endert

Co-organizer

Senior Consultant, DW Akademie (Germany)

Read biography

Lalya Gaye

Co-organizer

EBU Lead for vera.ai project

Read biography

Asmelash Teka Hadgu

Co-founder and CTO, Lesan AI (Germany)

Read biography

Kristophina (Kiito) Shilongo

Senior Tech Policy Fellow, Associate Researcher, Mozilla Foundation, Research ICT Africa

Read biography

Julie Ricard

Technology and Democracy Director, Data-Pop Alliance (Brazil)

Read biography