What makes you a strong team?
We combine diverse perspectives on the topic of AI and conduct research at the intersection of social science and AI development. Beyond AI Collective brings together developers and scientists to make AI as non-discriminatory as possible. Liquid Democracy tests and researches the interplay of democracy, participation and technology. Even while working on the draft, we noticed how well our competencies complement each other.
Explain your idea in three sentences.
With KI-MoDi, we want to use the example of AI-supported moderation of online discussions to show how an AI model can be examined for the risks of discrimination, from training the data to its practical application. We develop suggestions on how to make these risks transparent and counteract discrimination. With kosmo-moderation.de, we have already developed an AI moderation assistant and conducted initial laboratory tests, which we are examining in more detail in KI-MoDi.
What makes your idea special?
Our approach is holistic. This is why our idea of identifying and avoiding discrimination risks involves a mix of technical and non-technical measures. Thanks to our interdisciplinary team, we are able to influence the social sciences, software development and the community around digital democracy. Our results are freely available; they can and should be used further. As such, we hope to have a broad impact as a model project.
What’s next?
At a kick-off meeting, we will begin the detailed planning and fine-tuning of the objectives for KI-MoDi. We are very much looking forward to using the prize money as a team to intensively refine our concept for identifying and preventing models of discrimination over the coming months. We’re also curious about what other teams are planning and are looking forward to sharing ideas and working together.
The Policy Lab of the Federal Ministry of Labour and Social Affairs