With the exponential growth in IoT devices, especially in healthcare systems, where multiple devices are attributed to one person, the need for a secure and effective group secret key (GSK) generation is intrinsic. GSKs facilitate simplified and energy-efficient key management. Moreover, GSKs are suitable for encrypting broadcast control and alert messages. Existing group key generation schemes show many pitfalls, with the majority of them being either computationally inefficient or impractical. In this paper, we propose a novel approach to generating GSKs by exploiting the channel state information (CSI) of spatially correlated IoT nodes. We implement a feedforward neural network (FNN) for identical channel feature generation for the nodes. We then employ a group key generation scheme based on the generated channel features. Results verify the performance of the proposed scheme and demonstrate outstanding performance in the presence of a passive adversary.