The Dall-E mini has become a whimsical sensation over the past few weeks. Users are feeding in hilarious prompts to create random, blurry, funny images. Although the imaging libraries for AI are vast, it has started showing some limitations.
What happens when you give no guidance or direction to the AI? Write absolutely nothing in the search bar, and the AI model seems to get stuck.
Why is DALL-E mini obsessed with saris?
A report from Gizmodo says that with no prompt provided, the program will give you back an image of a brown-skinned woman wearing a sari (a garment commonly worn in South Asia, especially INDIA). This weird phenomenon makes us think, what is DALL-E mini’s bizarre obsession with saris?
A different report from Rest Of World says Marés, a veteran Hacktivist who has been using DALL-E mini since early June, came across this odd set of results every time he left the search bar blank. He repeated this process hundreds of times, and every time he ran a blank request, it was women in saris all over again.
Although it’s still hard to answer why the DALL-E mini is obsessed with this specific type of image, “it may have something to do with incomplete data sets and tagging,” the report said. Even the AI’s developer, Boris Dayma, doesn’t know exactly why it’s happening. “It’s quite interesting, and I’m not sure why it happens,” he said, according to Rest Of World’s report.
Even a little flawed, we should appreciate the presence of such a powerful AI image generator for public use. Unlike its older brother, the DALL-E 2, the DALL E mini has been open for public use since July 2021.
DALL-E 2 with open AI pushes the boundaries of an AI image generator by leaps and bounds. It produces impeccable realistic-looking images with whatever prompt is given. However, it’s still a hit-and-miss situation depending on the detail provided in the prompt.
Why do you think DALL-E mini is obsessed with saris? Comment down below.