Artificial Intelligence (AI) has significantly permeated diverse sectors, and one of its intriguing applications is observed in the creation of virtual headshots. These AI headshot generators employ Generative Adversarial Networks (GANs) to produce remarkably realistic images of non-existent individuals. As we delve into the future of this intriguing application of AI, it is essential to understand the underlying mechanisms, future trajectories, and potential societal implications.

At the heart of AI headshot generators are GANs, a class of machine learning frameworks designed by Ian Goodfellow and his colleagues in 2014. GANs comprise of two parts: a generator network, which produces synthetic data instances, and a discriminator network, which evaluates these instances for authenticity. Through this simultaneous process of generation and evaluation, AI headshot generators are able to create compellingly realistic images.

Emerging trends indicate that the sophistication and application of AI headshot generators are set to proliferate further. One of the remarkable advancements is in the area of "StyleGANs", a variant of GAN, introduced by NVIDIA in 2018. StyleGANs manipulate the 'style' — the color scheme, texture, and overall aesthetic — of the generated images, thus offering a wider range of customization.

Moreover, recent developments point towards increasing temporal consistency in AI-generated headshots. Currently, most AI headshot generators produce still images. However, integrating Recurrent Neural Networks (RNNs), a type of AI designed to handle sequential data, could pave the way for generating consistent sequences of images, potentially leading to the creation of AI-generated videos of non-existent people.

The prospective use-cases for AI headshot generators are as diverse as they are exciting. For instance, in the entertainment industry, these generators could create numerous extras for films and video games, reducing production costs significantly. Moreover, in the area of online privacy, AI-generated headshots could provide anonymous yet personalized avatars for social media users.

However, it is imperative to note that these advancements are not without concerns. The primary ethical issue arises from the potential misuse of AI headshot generators in creating deceptive imagery, known as 'deepfakes'. These manipulated images and videos can be used to spread misinformation and propaganda, posing significant challenges for digital media ethics and law enforcement.

From a broader societal perspective, the increasing realism of AI-generated headshots may also contribute to blurring the line between reality and virtuality. The philosopher Jean Baudrillard's concept of 'hyperreality', where the distinction between the real and the artificial becomes indistinguishable, seems increasingly relevant in this context.

In conclusion, while the future of AI headshot generators promises intriguing advancements and diverse applications, it also calls for robust ethical guidelines and regulatory frameworks. The challenges posed by this technology echo the sentiments of the renowned mathematician and philosopher, Alfred North Whitehead, who said, "The major advances in civilization are processes that all but wreck the societies in which they occur." As we stand at the brink of another major advance, it is crucial to ensure that the society harnessing it is adequately equipped to navigate its complexities.