Google Works to Improve Skin Tone Bias in Image Search Results

Google has announced the release of a new skin tone scale to make search results more inclusive.

The tech giant has teamed up with Harvard professor and sociologist Dr. Ellis Monk who has come up with a more equitable skin tone scale. The Monk Skin Tone (MST) Scale is designed to replace outdated skin tone scales skewed towards lighter skin.

Monk has been studying how skin tone affects the psychology of communities when faced with taboo topics such as colorism and underrepresentation of less celebrated tones in conditional settings.

“In our research, we found that a lot of the time people feel they’re lumped into racial categories, but there’s all this heterogeneity with ethnic and racial categories,” the 10-year veteran said during a conference call with UrbanGeekz

“And many methods of categorization, including past skin tone scales, don’t pay attention to this diversity. That’s where a lack of representation can happen…we need to fine-tune the way we measure things, so people feel represented.”

A part of that representation lies in the hands of the Monk Skin Tone Scale. After viewing results with participants nationwide, research demonstrated that the scale’s representation aligned with the current technology standard, especially for darker tones.

Google will use the MST Scale as a catalyst to grow an organic design for easier and more accepting recognition with consumers. The scale will also help the tech industry build more representative datasets to train and evaluate artificial intelligence models for equity in face imaging.

The MST Scale and supporting technology will improve how users use Google with better expectations. So for example, looking up images regarding makeup, the scale will optimize its search patterns to better match users relative to their interests.

The benefits of the tech instrument will also serve creators, brands, and publishers. They’ll be able to use this new inclusive schema to label their content with attributes like skin tone, hair color, and hair texture.

This way, they’ll be able to label their images with an SEO language easier for search engines and other platforms to comprehend.

During the call, a concerning matter was brought up revolving around darker tones and how the scale builds a relationship between sensitivity awareness and tech.

Monk drew attention to the Fitzpatrick Scale which is a numerical scale that measures the amount of melanin in the skin after exposure to the sun. The scale was developed to measure how skin reacts to ultraviolet light or sun exposure.

“The reality is the Fitzpatrick Scale didn’t do a good job differentiating darker skin tones which leads to the purpose of the MST Scale,” he said. “It’s intended to fix those kinds of issues creating a more dynamic range which fits comfortably among the Black population.”

Tulsee Doshi, who is the Head of Product for AI and Product Inclusion in Search, also commented on skin tone spectrums and how it’s received by consumers when using apps alongside Monk.

Doshi added, “This is why we are excited about the scale. Every user is different and deserves a good experience and you want to build off that which works so we won’t personalize but instead refine it.”

“We are taking the steps forward with the schema to create more relevant and beautiful imagery to make the experience better. But we’re at the beginning so we’re still building towards that sensitivity,” she said.

What’s next for the MST Scale? Google will continue collaborating with Monk to test the scale across different regions and product applications to enhance possibilities and inclusiveness.

• Tags:
• Artificial Intelligence
• Google
• Google Pixel
• inclusive technology
• Skin

COMMENTS.

comments