I threw a few very poor (they were shot on a Sony Mavica diskette-based camera) photos of me from 2000/01 and said, “Make photorealistic.”
I’m kind of surprised at the results. As you can see, some are certainly better than others. But like all AI generated images of humans, they still suffer from the “uncanny valley.” It’s art to be sure, and as good as they are, no one is ever going to mistake these for real photographs. There’s just something off about them.
I then asked ChatGPT to describe the person in these photos:
The man in these images appears to be a middle-aged adult with a friendly and approachable demeanor. Here are some consistent traits and observations based on the visual cues across the photos:
-
-
-
-
Facial Features: He has a round face, expressive eyes, and prominent eyebrows. His expressions vary slightly, but overall he maintains a warm and slightly amused look, often accompanied by a smile or smirk.
-
Facial Hair: The first image shows him with a goatee, while the others depict him with a full, thick mustache. This suggests some variation in grooming over time or for different styles/contexts.
-
Setting: All images appear to be taken indoors, in casual home settings — often with an aquarium in the background, indicating it might be his home or a familiar space.
-
Style: His attire is casual (T-shirts, flannel, casual button-downs), suggesting a laid-back personality. The “Roswell, NM” alien-themed shirt hints at a sense of humor or an interest in pop culture, sci-fi, or Americana.
-
Mood/Expression: Across the board, his expressions are relaxed, slightly playful, and self-assured. One photo features a tilted head resting on his hand, which gives a thoughtful, personable vibe.
-
-
-
Oh my.