Page 457 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

2 Use Case Description

2�1 Description

Use-case description:The case tackles the critical issue of equitable access to Artificial 4.4-Productivity
Intelligence (AI) tools for media organizations worldwide, in Eastern DRC in particular. As
AI rapidly transforms journalism and the information ecosystem, a significant digital divide
is emerging, leaving media outlets using low-resource languages at a disadvantage. The
case brings together media practitioners and technology to explore the challenges faced
by these media, with the examples of DRC, highlighting the lack of data in Swahili, suitable
algorithms, and technical resources. We will showcase innovative solutions, particularly from
African-led initiatives, and discuss the potential for AI to empower these outlets, focusing
on automated transcription and translation. What does African media need to be able to
leverage this transformative technology? How can we ensure that the rich diversity of languages,
cultures, and local knowledge are embodied in the process? Our goal is to foster dialogue and
collaboration, urging greater investment and support for a more inclusive media landscape.

Partner

Prof. Isabel Günther - Development Economics - ETH Zurich - Clausiussstrasse 47, 8092 Zürich
Partner url: https:// eth4d .ethz .ch/

2�2 Benefits of the Use Case

Access to information and the role of media play a key role to secure peace and justice in any
society. Yet, when local journalists from low resource languages and with limited access to new
technology can't work in the language the majority of their audience speaks, then there is an
inequality, and a gap. Bridging that gap for more peaceful and just societies is important and
can be done if local media in developing countries can access new tools and if local populations
can access information in the language they speak and in the media they locally use.

2�3 Future Work

Data collection, Create new variations/extensions to the same use case, Standards development
related to the use case Elaborate proposal: The case is currently monitoring 10 radio stations in
Eastern DRC, in Swahili and French. The transcripts of the content is done by AI, and the search
for disinformation or problematic content such as lies or hate speech is done by the machine
based on key words and knowledge provided by the local journalists.

We need to improve the speech to text in Swahili with more radio stations' contents, as well as
constantly update the trends to monitor as the situation in Eastern DRC is volatile and fragile.

3 Use Case Requirements

421