It would be an understatement to say that ChatGPT’s newly acquired native image generation capabilities have become one of its most popular features. The built-in image generator has already powered several viral trends—whether it’s turning real-life photos into Studio Ghibli-inspired art or rendering people as action figures. But now, social media users have taken it a step further: they’re asking ChatGPT to create mugshots of themselves—complete with absurd arrest reasons.
A trend gaining traction on Reddit involves users prompting ChatGPT to generate mugshots based on their usernames. But these aren’t just any mugshots, instead ChatGPT is also asked to make up bizarre reasons for the arrest, giving each image a funny twist.
Since then, a number of hilarious ChatGPT-generated images have appeared, each explaining the reason for the user’s arrest. For the username Nagadavadia, ChatGPT generated a picture of a white man with a beard and long hair, wearing glasses and a lizard-like creature sitting on his lead. The mugshot, while revealing the reason for the arrest, reads,“Charged with proclaiming himself the high priest of lizardkind”
Another user, Darthpug93, received a generated image of a man in a black t-shirt, wearing a Darth Vader helmet with dog ears and dark makeup, holding a sign that reads: “Charged with attempting to use the Force on a dog”
Why is ther a sudden craze for ChatGPT generated images?
ChatGPT has had the ability to rely on DALL-E 3 to convert text prompts into images for some time now. However, a few weeks ago, OpenAI unlocked the native image generation capabilities of GPT-4o (the language model that runs the chatbot), allowing it to create images without relying on an external model.
ChatGPT blends its image generation abilities along with its text knowledge base and the ‘memory’ of user’s preferences to create more nuanced and accurate images that often end up matching the demands of users.
Earlier this week, OpenAI also rolled out its o3 and o4 mini reasoning model which have since then been used on social media to identify locations based purely on visual cues in a photo.
Source link