Google Promises To Fix Its AI Image Bot After It Was Accused Of Being Too Woke

Google's AI chatbot was recently criticized for trying too hard and over-correcting prompts to avoid the risk of being racist.

The accusations aren't based on a single incident. In fact, a bunch of users have shared their experience and shown how the image results are often historically inaccurate. Some users have also mentioned that the bot seems to try and correct years-old racial bias all by itself.

For example, when it was given a prompt to generate an image of the founding fathers of the US, it produced an image of women and people of color in return.

Google's Response To The Incident

Google has acknowledged the issue and responded with a tweet on X which said they were working on this problem. The tweet also said that Google Gemini has been designed to create images of a wide variety of people because the tool is used in different corners of the world.

That's a good thing and that's exactly what it was trying to do here. But in this case, it was missing the mark".

Since historical prompts heavily depend on the context, they have to be treated with extra care.

The incident was also addressed by Jack Krawczyk, senior director for Gemini experiences at Google who said that they are now aware of the inaccurate historical depictions the tool is making and are working round the clock to fix it.

He said that the company takes the representation of various segments of society very seriously and is trying hard to minimize bias based on color, gender, and race. While this is a noble approach, in this case, it backfired.

He further added that while they'll continue this approach for open-ended prompts (such as a lady wearing a hat and making pancakes), historically prompts will be treated differently.

He ended his take on the incident by thanking the users for the feedback. Since Gemini is still very new to the market, this feedback will definitely help them finetune the tool.

What Do People Have To Say About This Incident?

While some users are happy with the statements, critics have a different opinion. Some even went on to say that the tool is too woke" and trying too hard to be politically correct and prevent accidental racial bias. We also noticed some friendly fire from its competitors such as Elon Musk.

I'm glad that Google overplayed their hand with their AI image generation, as it made their insane racist anti-civilizational programming clear to all.Elon Musk

It seems like the fear of accidental racism arose from a prior incident that led people to heavily criticize the tech company.

For instance, decades ago the Google photo app accidentally labeled the photo of a black couple as Gorilla". The incident not only brought racism accusations to the company but some were also left to question the kind of images the company was using to train its bots and if they were diverse enough to represent different categories of the population.

While Google has promised to make amends, they have not shared any timeline on when we can expect to see the results.

The post Google Promises To Fix Its AI Image Bot After It Was Accused Of Being Too Woke appeared first on The Tech Report.