Microsoft rolls back its Bing Image Creator model after users complain of degraded quality

Date:

Share post:


Ahead of the holidays, Microsoft said it was upgrading the AI model behind Bing Image Creator, the AI-powered image editing tool built into the company’s Bing search engine. Microsoft promised that the new model — the latest version of OpenAI’s DALL-E 3 model, code-named “PR16” — would allow users to create images “twice as fast as before” with “higher quality.”

But it didn’t deliver. Complaints quickly flooded X and Reddit.

“The DALL-E we used to love is gone forever,” said one Redditor. “I’m using ChatGPT now because Bing has become useless for me,” wrote another.

The blowback was such that Microsoft said it’ll restore the previous model to Bing Image Creator until it can address the issues.

“We’ve been able to [reproduce] some of the issues reported, and plan to revert to [DALL-E 3] PR13 until we can fix them,” Jordi Ribas, head of search at Microsoft, said in a post on X Tuesday evening. “The deployment process is very slow unfortunately. It started over a week ago and will take 2-3 more weeks to get to 100%.”

So what went wrong?

It’s difficult to compare model outputs from anecdotal reports, particularly when the prompts aren’t standardized. But many users said that PR16 tended to make images look less realistic — and “lifeless.” Mayank Parmar, writing for Windows Latest, noted that PR16-generated images lacked detail and polish, and appeared weirdly cartoonish.

It’s not the first time an image model that presumably passed internal checks wasn’t well received publicly. Back in February, Google was forced to pause its AI chatbot Gemini’s ability to create images of people after users complained of historical inaccuracies.

The missteps illustrate just how challenging it can be to measure model improvements in the real world. According to Ribas, Microsoft’s benchmarking found PR16’s quality to be “a bit better on average” compared to the previous Bing Image Creator model.

Whatever internal metric the company used, it seems clear that it didn’t align with most people’s preferences.





Source link

Lisa Holden
Lisa Holden
Lisa Holden is a news writer for LinkDaddy News. She writes health, sport, tech, and more. Some of her favorite topics include the latest trends in fitness and wellness, the best ways to use technology to improve your life, and the latest developments in medical research.

Recent posts

Related articles

Trump ends legal battle over Twitter ban

President Donald Trump’s lawsuit against the company formerly known as Twitter appears to be over. Trump sued the...

Apple’s new research robot takes a page from Pixar’s playbook

Last month, Apple offered up more insight into its consumer robotics work via a research paper that...

Elon Musk said he’s not interested in acquiring TikTok

Elon Musk recently said he is “not chomping at the bit to acquire TikTok.” Musk made those remarks...

Ex-Meta employee sues for sexual harassment

Welcome back to Week in Review. This week we’re diving into Google quietly removing its pledge to...

AI pioneer Fei-Fei Li says AI policy must be based on ‘science, not science fiction’

Fei-Fei Li, the Stanford computer scientist and startup founder sometimes known as “the Godmother of AI,” has...

Autodesk CTO Raji Arasu calls for diversity in the teams building AI

To give AI-focused women academics and others their well-deserved — and overdue — time in the spotlight,...

Why the ‘spirit’ of open source means much more than a license

Arguments about what is and isn’t “open source” are often resolved by deferring to the Open Source...

Christie’s announces AI art auction, and not everyone is pleased

Fine art auction house, Christie’s, has sold AI-generated art before. But soon, it plans to hold its...