Alphabet’s Google is silently working on developing some alternative to the standard method of classifying skin tones. This comes at a time when more and more researchers and dermatologists are saying that the industry-standard method is not sufficient for assessing bias in products against people of colour. The tech giant hopes that the new standard for classifying skin tones would save it from embarrassing acquisitions of being biased in future. The development coincides with reports of numerous problems being reported because of how artificial intelligence and machine learning deal with different skin tones. The problem is extended to wearable and other devices. There is a spate of high-profile mistakes along with misclassification of Black people in cloud photo services. There are also incidents where smartwatches and wearables couldn’t take heart rate and missed other measurements from darker skin. Most of the tech companies have pointed that it is the result of inadvertent and unforeseen side effects of the systems. But Google has been silently working on finding a better way to deal with it. The company is working on finding a new way to handle the diversity of users. While Google has confirmed the news, it has refused to divulge any detail related to it.
“We are working to develop more inclusive measures that will help the development of products in a better way. We are looking forward to collaborating with medical experts and scientist along with groups that are working with communities of colour,” Google was quoted as saying in a report. The issue is mainly because of the six-colour scale currently being used. Known as Fitzpatrick Skin Type or FST, the scale has been around since the 1970s and widely used by dermatologists. The scale categorizes skin colour into six tones. FST includes four categories for ‘white’ skin people, one for ‘black’ and one for ‘brown’ skin people. Most of the tech companies have been using this scale to categorise people. But critics are of the view that FST disregards diversity among people of colour.
The FST which are currently being used was developed to classify skin types by which these people were prone to sunburn. These groupings were then used by medical experts for the treatment of ultraviolet radiation. But the scale soon gained popularity and more several companies adopted it as a standard to segment groups of people while developing products. Even emojis widely used on popular social media platforms used FST for skin tone selection. The tech giant has not fixed any deadline about when its alternative system would be ready for deployment. It is also not clear whether the company will silently deploy the new system at the backend or will discuss about it publicly. But it is to be noted that no other tech companies seem to be taking this matter seriously and working on developing more effective and balanced scales.