homesmart tech NewsGoogle unveils 10 shade ‘Monk Skin Tone Scale’ to improve inclusivity

Google unveils 10-shade ‘Monk Skin Tone Scale’ to improve inclusivity

The 10-shade skin tone scale will be integrated into Google’s AI infrastructure and all imagery products

Profile image

By CNBCTV18.com May 12, 2022 12:59:25 PM IST (Updated)

Listen to the Article(6 Minutes)
Google unveils 10-shade ‘Monk Skin Tone Scale’ to improve inclusivity
Google has adopted a 10-grade skin tone scale that shows the spectrum of real-world skin tones to improve the representation of various skin tones across its products. The scale is called the Monk Skin Tone Scale, which has been developed in collaboration with Harvard professor and sociologist Dr Ellis Monk.

Dr Monk has been studying the effects of skin tone and colourism on people’s lives for more than a decade now. Google announced that the MST scale will be soon implemented across its AI infrastructure in all Google products.
What is the Monk Skin Tone scale?
The Monk Skin Tone scale is named after Dr Ellis Monk, who is a Harvard University professor and sociologist. He has spent decades researching colorism’s impact on the lives of Black people in the US. He created the scale in 2019 and has been working with Google engineers and researchers to incorporate it into the company’s product development.
However, the MST Scale does not encompass all possible skin tones, but as per Google, it is far more representative than that of the current tech industry standard.
Tech companies have been using the Fitzpatrick scale to classify skin tones for their computer vision algorithms. The scale was designed for dermatologists in the 1970s and had only six skin tones. This possibly contributed to the AI’s well-documented failures in identifying people of colour. Racist soap dispensers and autonomous driving systems labelling people with dark skin as less reliable are a few examples of the pitfalls of computer vision algorithms. One such failure of Google was a photo categorization feature of Google photos that labelled pictures of black people as gorillas.
Google’s team and Dr Monk surveyed thousands of adults in the US to learn if people felt more represented by the MST Scale compared to other scales used in both the machine learning and beauty industries.
“Across the board, people felt better represented by the MST Scale than the Fitzpatrick Scale and this was especially true for less represented demographic groups,” says Xango Eyeé, a product manager working on Responsible AI in Google’s blog.
How will the MST help in better representation?
As per Google, using the MST scale can help AI better understand representation in imagery so that it can be sure that a product or feature works well across a range of skin tones. This is particularly important to computer vision algorithms.
“The MST Scale will help us and the tech industry at large build more representative datasets so we can train and evaluate AI models for fairness, resulting in features and products that work better for everyone of all skin tones” the company wrote in its blog.
Google aims to use the MST scale to better provide skin tone representation in Google Search results. For instance, if someone searches for makeup-related queries, they would be able to refine results to show images by skin tone.
Google Photos will also be integrated with the MST Scale. Last year, the company added a new way to improve its auto-enhance feature with its Real Tone filters for the pixel that is designed to work well across skin tones and now they will be evaluated using the MST Scale. Google has worked with a diverse range of renowned image-makers like Kennedi Carter and Joshua Kissi, who are celebrated for beautiful and accurate depictions of their subjects, to evaluate, test, and build these filters.
Real Tone filters will also be released on Google Photos across Android, iOS, and the Web in the coming weeks.
Google has also released the MST scale for anyone to use to improve products and provide feedback to further develop the system. Google is further continuing work to improve the Monk Skin Tone scale in places like Brazil, India, Mexico, and Nigeria, as per a WIRED report.
 

Most Read

Share Market Live

View All
Top GainersTop Losers
CurrencyCommodities
CurrencyPriceChange%Change