Historically, although this is still a very current problem, cameras have used technology that does not take into account all the tonal diversity of skin that exists in the world, always favoring the light tones of whites. Racism in all its glory.
At Google I/O last year, the company showed off “Real Tone,” a new feature on Pixel 6 phones meant to provide a better photography experience for people with darker skin tones.
In this year’s edition, Google has taken a step closer to full social representation with the presentation of new initiatives that finally use technology to fight this systemic inequality.
“Google must meet the diverse search needs of billions of people. And it’s critical that they’re reflected in our products,” said Prabhakar Raghavan, senior vice president at Google.
Let’s see how Google will make its products more suitable for non-white people.
The Monk Skin Tone Scale
To bring about all these changes, Google collaborated with Harvard University sociologist and professor Ellis Monk, who has spent more than a decade studying how skin tone and colorism can affect people’s lives.
Monk is the author of the so-called Monk Skin Tone Scale, also known as MST, a scale of 10 shades that Google is now introducing into some of its products to show its commitment to social representation and equity in the world of beauty too. Photography.
The artificial intelligence and machine learning used today can “perpetuate unfair biases and may not work well for people with darker skin tones.” That’s because they weren’t designed to “see” and “understand” those skin tones.
Hopefully (and a lot of will on the part of the technology industry), the scale that Google presented, open source, will be used by other companies in the sector to also use it in their products and so broaden the spectrum of skin tones.
It is, as we said, a 10-tone scale, ranging from lightest, #f6ede4 (HEX), to darkest, #292420 (HEX). To make it easier for you, Google has a new web page where you can find all the information you need to know how to use it.
“Monk Skin Tone Scale” by Dr. Ellis Monk, licensed under Google LLC under Creative Commons Attribution 4.0 International License
Google searches
“When you look for beauty tutorials, you only see white women. And that contributes to the fact that they are not represented”, explains a testimonial that we could see in the presentation of the Google I/O.
“Being represented is a right. If the internet is really for everyone, then everyone should be represented,” said another.
If Google really wants to “make products for the world,” it needs to start by reaching out to non-white people who don’t feel included by even the most popular search engine. The company will start using the Monk scale to change it.
Improvements to Google Images will help find results that show a wider range of skin tones. This will happen both when we search for a more generic topic like “people”, but also in more specific searches like “makeup ideas”.
Google Images users will be able to use a new complexion filter so that the results displayed by the browser are more precise, relevant and representative of their reality.
To deliver these results, Google relies on the collaboration of content creators in line, as they need their content to be well labeled for the new technology being implemented to work effectively. In the coming months, the company will release a new standard for assigning the correct label to skin tone or hair color and texture.
Google Photos
At the end of May, the company will launch new “Real Tone” filters that should work well with all skin types, as they were rated using the MST scale. They have also collaborated with photography professionals to create these filters to improve the rendering.
“With Real Tone Filters, you can apply the beauty and authenticity of professional editing to your own photos with just a few clicks,” said Annie Jean-Baptiste, Head of Inclusion and Equity at Google, at Wednesday’s event. And now that will be true for non-whites as well.
The new Google Photos filters will be available for Android and iOS devices and on the web.