Prof. Dr. Louisa Specht-Riemenschneider, Federal Commissioner for Data Protection and Freedom of Information, and Prof. Dr. Matthias Braun, Chair of (Social) Ethics, captivated the young audience with their engaging presentations. The event took place in the Wolfgang-Paul-Hörsaal.
The children participated enthusiastically, eager to share their knowledge on topics such as artificial intelligence, data protection, and the use of mobile phones. With the children's active support, Louisa Specht-Riemenschneider began by explaining what data actually is: Data consists of statements or information about us that can be found on the internet. To prevent our data from ending up unfiltered on the internet, we have to protect it–e.g.,by choosing secure passwords or ajusting privacy settings.
The lecture continued with two examples of how data is used and what effects this can have. With the help of two different prompts to an image AI, the children were given an impressive demonstration of how data can reproduce prejudices: An AI draws on data that exists on the web or that it is fed with. For example, if the prompt “A woman is sitting on a horse” is entered, the AI accesses the data that exists on the web. In this way, it reproduces ideals that exist in our society. In the case of the first prompt, the AI-generated image showed a white woman who looked “like a Barbie” to the children. Through constant repetition, AI can even influence what we see as normal or abnormal. The second example was the prompt “A child is walking along a dirty path”. Here, the children noticed that the children in the pictures looked different from the woman on the horse: the children affected by poverty were depicted with darker skin.
Using these two examples, the children realized that data can (re)produce prejudices, also known as bias. Together with Prof. Dr. Specht-Riemenschneider and Prof. Dr. Matthias Braun, the children came to the conclusion that although artificial intelligence can be used to do fun things, it can also have dangerous effects when it comes to reproducing discrimination and marginalization or shifting norms or prejudices. This means that we need rules to define what is allowed and what is not. One proposed measure to prevent damage is labeling AI-generated videos or photos to ensure transparency.
In the end, the children applauded and tapped on the tables. After the lecture, they were given the opportunity to ask further questions, get autographs or take photos with the speakers.
The Children's University is a series of events organised by the Junge Uni Bonn. You can find more information about the programs for schoolchildren here1.
