Summary

Synthetic data can be a powerful tool in research and development, but it also comes with challenges. In this webinar, Professor Ericka Johnson (Linköping University) introduces key concerns around bias in synthetic datasets, present tools (which are freely available to SND DORIS users) for bias-assessment, and discusses how metadata can help make such biases and other generative “hallucinations” more transparent.

Description

The seminar will close a presentation of new policy proposals from the AI Policy Lab, and insights from the September 2025 workshop in Norrköping.

Note that this webinar will be in English.

When: October 7, at 11:00-12:00

Where: Zoom (link to this webinar on Zoom) 

Contact: David Rayner david.rayner@snd.se