This discussion paper results from ongoing discussions with representatives from different platforms on national and European level on how a human centric metric of trust in mixed human-AI settings in Industrial AI could be developed. With growing complexity, AI systems may suggest solutions humans have not thought of – leaving the human puzzled. Being surprised about the unexpected, the human will want to find out the reasons behind to gain more insight, starting a dialogue. The ZVEI working group “Artificial Intelligence in Industrial Automation” proposes the “levels of surpisedness” as the metric and a dialogue based interaction method to gain trust in mixed human-AI settings.
We invite you from industry or academia, being supplier, integrator or end user to join in for further discussions!