Can AI, Data and Robotics Research be Open and Accessible to All? Exploring the challenges From the perspective of open research data management

The application of open and FAIR data principles remains a challenge to this day. Besides the technical challenge that those new work habits represent, biases in research, such as a positivity bias or a global north bias, cannot be properly addressed by open data alone. When those datasets end up as training material for AI R&D, they provide a skewed picture of reality that prevents AI developments from being completely trustworthy. What can research stakeholders do to address those biases in open data?