Content of the article
AUCKLAND – Google’s Alphabet Inc on Wednesday unveiled a palette of 10 skin tones, which it described as a step forward in creating gadgets and apps that better serve people of color.
The company said its new monk skin tone scale replaces the flawed six-color standard known as the Fitzpatrick Skin Type, which has become popular in the technology industry to assess whether smartwatch heart rate sensors, artificial intelligence systems, including facial recognition and other suggestions show color bias.
Content of the article
Technology researchers have admitted that Fitzpatrick did not represent people with darker skin. Reuters exclusively announced last year that Google was developing an alternative.
The company has partnered with Harvard University sociologist Ellis Monk, who studies color and has felt dehumanized by cameras that fail to detect his face and reflect the complexion of his skin.
Monk said Fitzpatrick is great for classifying the differences between lighter skin. But most people are darker, so he wanted a rock that “does a better job for most of the world,” he said.
Monk through Photoshop and other digital art tools selected with 10 tones – a manageable number for people who help with learning and evaluation of AI systems. He and Google surveyed about 3,000 people in the United States and found that a significant number said the 10-point scale matched their skin, as well as a palette of 40 shades.
Content of the article
Tulsi Doshi, product manager for Google’s responsible AI team, called the Monk scale “a good balance between being representative and bearable.”
Google is already implementing it. Beauty-related searches on Google Images, such as “bridal makeup look,” now allow you to filter results based on Monk. Searching for images like “cute babies” already shows photos with different skin tones.
The Monk Scale is also being implemented to ensure that many people are happy with the filtering options in Google Photos and that the company’s face matching software is not biased.
However, Doshi said problems could spill over into products if companies did not have enough data on each tone or if people or tools used to classify others’ skin were biased by differences in lighting or personal perceptions.
(Report by Paresh Dave; Editing by David Gregorio)
Add Comment