Dr. Jessica Szczuka is the Head of INTITEC and a leading researcher in digitized intimacy. She holds a Bachelor’s and Master’s degree in Applied Cognitive and Media Science from the University Duisburg-Essen, where she also earned her Ph.D. in Social Psychology: Media and Communication.
Throughout her doctoral studies, Dr. Szczuka dedicated her research to empirically exploring the influence of digitalization on interpersonal relationships and sexual experiences.
Building Calibrated Trust in AI Event
Thematic focus:
In an increasing number of application fields (e.g., medical decision support systems, AI algorithms) it is necessary to guarantee „calibrated trust“. Most current AI systems do not provide calibrated trust, i.e. they are frequently characterized by a mismatch between perceived trustworthiness and actual reliability of the intelligent system, leading to its misuse (using the system even though it is not trustworthy) or disuse (not using the system though it is highly reliable and actually helpful). In order to calibrate human trust we need to develop AI systems which are a) more trustworthy than current systems and b) communicate a degree of trust which actually matches the capabilities (e.g., with self-explainable, accountable ML algorithms). For this, we need to develop and apply new methods to specify requirements for trustworthy systems and their verification, quantify uncertainty, specify formal and statistical reliability guarantees, preserve privacy and security and empower humans to understand, control, and verify intelligent algorithms.
