Purpose: To analyze how AI represents older adults, identifying potential stereotypical patterns using objective visual content analysis and perceptual dimensions of valence and dominance. Methods: Eight images were generated by AI using the prompt: “Create a hyperrealistic image of an 85-year-old woman and man.” The images were analyzed across eight dimensions: facial expression, posture, environment, clothing, color palette, dominant stereotypes, ethnic diversity, and perceptual variables (valence and dominance). Results: All images (100%) depicted Caucasian phenotypes, with neutral or slightly sad expressions and static postures. According to the Oosterhof Todorov framework, the images predominantly occupied quadrants of low valence and low dominance, associated with vulnerability. Conclusion: AI generates homogeneous representations of aging with low emotional and ethnic variability. The use of validated perceptual models allows evidence of how AI reinforces ageist stereotypes under the guise of aesthetic neutrality.