Tech companies have used the Fitzpatrick scale to classify skin tones for years. The system was originally designed for dermatologists and only has six skin tones. The Monk Skin Tone scale is a 10-skin tone standard that is being incorporated into all of the products of the company. The development has the potential to reduce bias in data sets used to train artificial intelligence.
Last year, there was a hint of plans to go beyond the Fitzpatrick scale. The project dates back to a summer 2020 effort by four Black women at Google to make artificial intelligence work better for people of color. At the I/O conference, the company showed how the new system could affect its products. Fitzpatrick is the industry standard for evaluating the fairness of cameras and computer vision systems, and it could be replaced by the MST.
There are images of people's faces being used where we need to test the algorithm for fairness, says Eye.
The Monk Skin Tone scale is named after Ellis Monk, a Harvard University sociologist who has spent decades researching colorism in the United States. Monk created the scale and worked with engineers and researchers at the company to incorporate it into their product development.
The reality is that life chances, opportunities, all these things are tied to your genetic makeup, Monk said in a video shown at I/O. I think this is a huge step forward.
The initial analysis by Monk and Google found that participants felt better represented by the MST than by the Fitzpatrick scale. According to an FAQ published Wednesday, having more than 10 skin tones can add complexity without extra value, unlike industries like makeup, where companies offer more than 40 shades. According to a source familiar with the matter, the Monk Skin Tone scale is still being worked on in places like Brazil, India, Mexico, and Nigeria. There will be more details in an academic research article.
The company will use the MST more. The option to sort makeup-related search results by skin tone based on the scale will be available later this month. The 10-skin-tone scale could have implications for the evaluation of search results, as well as other products.