AI-Generated Images in Publications: Navigating Disclosure and Consumer Rights

A recent incident involved a reader purchasing a professional photography guide, only to discover that many of the images featured AI-generated characteristics. These included distorted facial features, illogical lighting, and anatomical impossibilities such as models with six fingers or toes. The publisher acknowledged that the author failed to disclose the use of AI-generated images, and the review process did not catch these anomalies. In response, the publisher offered unconditional refunds to affected consumers, highlighting a failure in quality control and disclosure.

To address the growing concerns surrounding AI-generated content, China implemented the 'Measures for the Identification of Artificial Intelligence Generated and Synthesized Content' on September 1, 2023. This regulation mandates that AI-generated content must be clearly and conspicuously labeled. The aim is to ensure that consumers are aware when they are interacting with AI-created material, thereby protecting their rights and maintaining the integrity of information.

Intellectual property lawyer You Yunting emphasizes that for knowledge-based books that rely on accuracy and factual representation, such as photography guides, the use of AI-generated content requires even greater caution. From the perspective of the "Law of the People's Republic of China on the Protection of Consumer Rights and Interests," consumers are entitled to the right to know. Both authors and publishers bear responsibility. Publishers and editors are expected to act as gatekeepers, ensuring that works presented as human creations are indeed so, or that the use of AI is clearly communicated to readers.

The 'Measures for the Identification of Artificial Intelligence Generated and Synthesized Content' require both explicit and implicit labeling methods for AI-generated images. Explicit labeling involves visible text on covers or near images, while implicit labeling uses digital watermarks embedded in image data. However, explicit labels can be removed, and creators using overseas AI platforms can bypass implicit identifiers. This necessitates increased diligence from publishers in their review processes. Zhao Zhanling points out that books serve as vessels of knowledge and carry reader trust. When AI prompt generation replaces dedicated creation, and editors struggle to discern AI content, the core value of the publishing industry is at risk. Maintaining transparency and ethical practices is crucial for the long-term credibility and sustainability of publishing in the age of AI.