For much of modern history, photographs have carried a special authority. We often treat images as evidence because they seem to show us what was really there. Generative AI complicates that assumption. Images can now be created with remarkable realism, while authentic photographs can also be dismissed as fake simply because they look unusual or politically inconvenient.
In Visual Credibility in the Age of AI, I start with a more fundamental question: what makes an image credible in the first place? Drawing on research in communication, psychology, journalism, and computer science, I examine how visual credibility is built, challenged, and sometimes restored. I argue that credibility has never come from the visual medium alone. It is a collective achievement, made possible by technological advancements, professional practices, and institutions that allow images to function as credible evidence.
The book traces several interconnected challenges that AI-generated media pose to visual credibility: deceptive images that mislead people, growing skepticism toward authentic images, the amplification of polarized discourse, and a flood of visually spectacular and emotionally engaging content that competes with authentic visual evidence. I also examine the increasing delegation of credibility judgments to AI systems and what happens when decisions about what is real or trustworthy are increasingly mediated by machines.
I then examine possible responses to these challenges. AI does not destroy visual trust. But it exposes how much that trust has always depended on the social and institutional systems around images. I consider how technology, journalism, platforms, media literacy, and public policy can help rebuild visual trust, as well as what institutional capacity is needed to govern an information environment in which synthetic and authentic images increasingly coexist.
This manuscript is currently under contract with Cambridge University Press.
YILANG PENG 