Reconfigurable intelligent surfaces (RISs) open the way to one of the most fascinating technological advances: passive transmission. RIS turns reflection from a source of interference to an entirely new way to transmit the wireless signal that is well suited to the dense, high-traffic, high-obstruction areas where the need for capacity is the highest and connectivity often disappointing. In this Sparring Partners with Melike Erol-Kantarci at the University of Ottawa, we talked about her research on how RIS can improve capacity and coverage by creating a sublayer of elements that are small, inexpensive and energy-efficient (passive transmission does not require power), how AI is used to tune RIS phase shift and which use cases RIS can support, and what is the roadmap from 5G to 6G. The challenge? Operators need to manage all the RIS units and integrate them within the macro infrastructure. If using mmWave – as it will be most likely – RISs will be densely deployed, thus increasing the complexity of coordination. This is where AI and ML are necessary: they make RIS manageable by optimizing the deployment and operations of RIS. Mel gave us an overview of her research in this area and of what we should expect from AI/ML-driven RIS.