This study investigates the relationship between semi-supervised learning (SSL) and open-set recognition (OSR) under the context of generative adversarial networks (GANs). Although no previous study has formally linked SSL and OSR, their respective methods share striking similarities. Specifically, SSL-GANs and OSR-GANs require their generators to produce samples in the complementary space, which are then used to regularise their respective classifier networks. In turn, classifiers trained under SSL and OSR generalise the open space by tightening classification boundaries around the labelled categories. In other words, a classifier trained using an SSL-GAN intrinsically achieves OSR and vice-versa. To prove this SSL-OSR link, we theoretically and experimentally compare the state-of-the-art SSL-GAN with the state-of-the-art OSR-GAN. Our results find that all SSL-GANs and OSR-GANs work towards the same goal, and that the SSL-optimised Margin-GANs set the new state-of-the-art for the combined task of SSL-OSR. Future studies could further explore the theoretical similarities between SSL-GANs and OSR-GANs, as well as extend SSL-OSR to other learning policies.