New Delhi:
The internet has found its latest obsession: the Ghibli trend. Thanks to OpenAI’s newest image-generation tool, social media users, including celebrities and politicians, are transforming their pictures into dreamy, Studio Ghibli-style animations. For one woman, though, the trend took an unexpected turn when ChatGPT gave her an extra hand.
She initially uploaded a picture of herself seated at what looked like a cafe, her face resting in both hands. When she used the AI tool to reimagine the scene, the image retained most of the original details – except for one peculiar addition. A third hand appeared in the artwork, holding an ice cream cone.
The AI also altered the setting. Instead of a cafe, she was now seated in an open field, with an ice cream shop in the background, as if it had always been part of the scene.
She posted the blunder on Instagram with the caption, “ChatGPT-Tumse na ho paayega..” (You’re not cut out for this). Shocked and funny, had to post it.”
The post quickly went viral.
A user commented, “Wait… what… why three arms and hands.”
Another commented, “The chair legs near the table behind you form a cone-like shape – almost like an ice cream cone, which is causing GPT to hallucinate.”
“ChatGPT knows there is goddess in every Indian woman,” someone wrote.
Just as users were laughing off the extra-hand blunder, another AI misstep surfaced – this one both hilarious and eerily unsettling.
This time, the AI mishap involved Chhath Puja, a festival in Bihar. A user asked ChatGPT to convert a photo of women performing traditional rituals into a Studio Ghibli-style illustration. The original image was of a group of women standing by the water, holding baskets filled with fruits, incense sticks, and coconuts as offerings to the God.
In its attempt to recreate the image, ChatGPT misinterpreted one of the coconuts as a human head. The final artwork showed a woman on the left, holding a basket that should have contained a coconut but instead carried what looked disturbingly like a severed human head.
OpenAI’s latest image generation tool in ChatGPT, powered by GPT-4o, allows users to create photorealistic and artistic images, including the viral Studio Ghibli-style trend. Initially, for premium users, it is now available for free with a daily limit of three images.
