![]()
Abstract: Synthetic data has long been viewed as a great opportunity for greater data collaboration, dataset augmentation, and amplification. However, artificial data also represents a threat to society, and the benefits should not be viewed in isolation from the potential risks of misuse, increased algorithmic bias, and data dilution. It is a paradox that we want synthetic information to be both indistinguishable and distinguishable from authentic content; a dichotomy that has prompted a virtual arms race of AI detection and watermarking technologies, opposed by methods to circumvent those same technologies. If we cannot reliably distinguish between synthetic and genuine data, we risk contaminating and diluting decades' worth of data collection.
It is at the same time an event from the Center of AI Ethics in the Data Science community:
https://www.sdu.dk/da/forskning/c-ai-ethics/news-and-events/event-digital-kessler-syndrome
