Google has tweaked technology to allow gadgets and apps to better serve people of colour.
The tech company, which is part of Alphabet Inc, said the 10-colour palette of its new Monk Skin Tone Scale replaces the Fitzpatrick Skin Type, a standard of six colours that had become popular in the tech world to assess whether smartwatch heart rate sensors, AI facial recognition and other systems showed a colour bias.
Google's new initiatives are aimed at improving product experiences for people of all skin tones, as tech researchers found Fitzpatrick underrepresented people with darker skin.
The new 10-point skin tone scale was created with Harvard University sociology professor Ellis Monk, whose research has focused on colourism and skin tone.
The scale was found to be more representative of a greater range of skin tones, with the company now releasing the set of colours and information about the scale so that others can use it or suggest improvements.
Google envisions the scale as a standardised method of building and testing products across the range of human skin tones, and will allow companies to discuss which ranges of colours are well suited to specific products.
Within Google, these new scales will likely be used to evaluate search tools, facial recognition algorithms, and other automated systems to ensure they perform well across a wide range of skin tones.
A variety of technologies, including some of Google's AI systems, to cameras have been criticised for performing better with images of white people than people of colour.
Google likely will annotate images based on which skin tones appear on the Monk scale so that it can train and test AI and other technology on a diverse set of skin tones as it moves away from that era, indicate reports.
Google is also rolling out new filters for Google Photos as part of its existing Real Tone system, designed to help generate high-quality images for a broad range of skin tones.
The company is also looking into ways for publishers to indicate skin tones where they're relevant, similar to how recipe publishers can now add meta data to help search engines and their users find food instruction with certain features.