Page 38 - AI for Good - Impact Report

AI for Good

Focus Area: Rising Inequalities

As AI continues to reshape industries and societies, ensuring that its benefits are distributed
equitably across different demographics and regions has become a critical concern. The
133
transformative potential of AI should not be reserved for the most privileged, but rather, it must
help to reduce inequality and foster social equity. It is also important to make sure that the Global
South can benefit from AI solutions in an equal manner to the Global North.

Governments should continue to find new development approaches that can include users
from all backgrounds, as well as drive collaboration with other countries to share best practices
or support where needed.

Established practice: FAIR Forward – Artificial Intelligence for All

Fair Forward is an initiative from the Federal Ministry for Economic Cooperation and Development
in Germany that is “dedicated to the open and sustainable development and application of
artificial intelligence and particularly supports partnering countries in Africa and Asia”.
134

This initiative focuses on three goals: 1) Strengthen local technical know-how on AI in Africa
and Asia, 2) Improve access to training data and AI technologies for local innovation, and 3)
Develop policy frameworks for ethical AI, data protection and privacy. FAIR Forward is already
active in five countries. Ghana, Rwanda, South Africa, Uganda and India, and is supporting
the fair development of AI in the Global South. One interesting use case is the development
in Rwanda of open AI training data sets in the local language, Kinyarwanda, to make sure that
AI-driven apps and solutions are accessible to all.

Emerging practice: Brazilian Artificial Intelligence Strategy

Brazilian Artificial Intelligence Strategy (EBIA), launched in 2021, represents an innovative
approach to integrating human-centered values and fairness into the national AI agenda.
135
This strategy is particularly focused on addressing social inequalities through the responsible
development and deployment of AI technologies. The EBIA’s ethics guidelines are designed
to ensure that AI systems do not reinforce existing social disparities but instead contribute to
greater inclusivity and fairness.

The EBIA recommends implementing a national framework for AI ethics, incorporating
review mechanisms for AI usage by both private and public entities. Additionally, it proposes
integrating features in AI public solutions to mitigate bias risks, such as setting a false positive
threshold for face recognition to minimize identification biases. Furthermore, the text advocates
for promoting transparency in AI-driven actions to ensure alignment with democratic principles
and human dignity.

Technical Feasibility and Infrastructure

Focus Area: Sustainability of AI

The significant energy consumption of AI systems poses serious environmental concerns,
particularly in the context of climate change. Training large AI models can require vast amounts
of electricity that usually do not come from renewable sources. Additionally, the impact of
136
AI on water and waste creation is also largely underestimated, as the underlying technology
requires important water quantity to cool down the servers, which are also challenging to

28