The field of text-to-image models is rapidly evolving, with a growing focus on improving the safety, equity, and sustainability of these models. Recent research has highlighted the importance of considering diverse human experiences and values in the development of text-to-image models, and has introduced novel approaches to align these models with human values. Another key area of research is the evaluation of the safety and accuracy of text-to-image models, with a focus on identifying and mitigating potential biases and harms. Notably, several papers have made significant contributions to the field, including the introduction of new datasets and frameworks for evaluating and improving the safety and sustainability of text-to-image models. Notable papers include: SustainDiffusion, which optimises the social and environmental sustainability of Stable Diffusion models. DREAM, a scalable red teaming framework to automatically uncover diverse problematic prompts from a given text-to-image system. From Seed to Harvest, a hybrid red-teaming method for guided expansion of culturally diverse, human-crafted adversarial prompt seeds.