On the Trustworthiness Landscape of State-of-the-art Generative Models: A Survey and Outlook