Societal Biases in Slovak AI

Artificial Intelligence (AI) is progressively becoming more integrated into our daily lives. However, there are serious questions about the fairness and equity of this technology concerning various groups of people.

At KInIT, we are deeply involved in this topic, and we strive to share the results of our project “Societal Biases in Slovak AI” with experts and the general public alike. The project is focused on gender biases that might be present within AI systems working with the Slovak language. Our team examined stereotypes and discriminatory forms of behavior that appear in machine translation, speech recognition, and currently popular language models.

We invite you to an event that will be held on Wednesday, October 18, 2023. We will present our main findings regarding Slovak AI systems and show what kind of gender biases are present within them and how strong they are. 

Researcher Miro Dudík will deliver a guest talk on the topic “Putting AI Fairness Research into Practice.” Miro Dudík holds the position of Senior Principal Researcher at Microsoft Research Lab. He earned his doctoral degree from Princeton University in 2007, and is author of the Fairlean toolkit that is used to evaluate and improve fairness of machine learning models. 

To conclude the program, there will be an expert panel discussion with Mariana Szapuová, Veronika Koišová, Štefan Oreško and Marián Šimko, in which you can also participate via Slido using the hashtag #AIpredsudky.

Agenda

The event will be conducted in Slovak language, except Miro Dudík’s lecture, which will be delivered in English.

You can watch the event recording here:

This project was carried out with the financial support from the the United States Embassy in Slovakia.