A research team in the United States has developed an innovative AI tool capable of evaluating both the survival risk and physiological age of cancer patients using facial photographs. Initially, the AI underwent pre-training on a dataset comprising 40 million facial images. Subsequently, it was refined using data from over 34,000 cancer patients, showcasing remarkable predictive accuracy when applied to new, unseen test cases. The two AI models, each serving a distinct function, focus on different aspects, and their combined application enhances overall precision. The specific facial regions analyzed by the AI differ, with some judgment patterns being interpretable, while the health indicators captured by others remain enigmatic to human observers. At present, this technology is being implemented in clinical pilot programs, and the general public is encouraged to contribute selfies to further refine the AI's performance. Nevertheless, this advancement also raises concerns, including the potential for misuse and ethical dilemmas.