First question is what kind of audience do you expect? Also multidisciplinary? Students? ERCs? Senior researchers? Managers? Funders? Publishers? I think the questions should ideally be questions the audience would like to ask but can't think of at that moment. Students and ERCs are quite different in their view on open science in my experience. Students seem more guided by idea of a logical and efficient system. ERCs tend to worry more about career effects. Questions on how to move towards a more open and efficient system while avoiding detrimental effects for ERC careers and not putting burden on new generation may be best to capture audience attention. Another starter could be to ask the panelists what open science means to them. Simply ask them all to write down a definition and then share that with the audience. The writing down is in order to not have them influence each other. You can even involve the audience in this. Another idea is to focus on how to integrate open science practices in graduate school courses if professors themselves have not yet adopted these practices. I also like the constructive approach, focussing on small things you can start doing next week and praising people for what they already do. Be prepared though to in the 2nd half of the session also address 1 or 2 more contentious issues like impact factors, scooping, costs etc. A final idea is to not only ask about the how but also the why: things like relevance (societal problems, output availability, engagement), rigour (efficiency, replicability, open data/code/methods) and equity (participation, contributorship). Or even make it more personal: "What goals do you want to achieve/contribute to by practising open science?". It may be interesting to hear a bit about individual drives/motivations.
Another idea is to work with statements. In our workshop for young researchers in neuroscience in Barcelona we used 12 statements. They are in slide 51-64 of this slidedeck: https://doi.org/10.6084/m9.figshare.5110873.v1.