Google changes the way to classify skin color, calmly works on the new system expected search giant will avoid embarrassing allegations in future tech. The ways in which machine learning and artificial intelligence, along with other devices, can be used and other devices, dealing with different skin colors has caused many problems in recent years.
It ranges from the error classification of black people in cloud photo services, durable and smartwatch that cannot take heartbeat and other measurements of darker skin, and a series of other high profile errors. While technology companies have protested their innocence, pointing to unintentional side effects – and unexpected from the system, work seems to have been going on on Google, at least, it is a new way to deal with user diversity.
The problem, it is recommended, is the color scale currently used, FitzPatrick or FST skin type. They have been around since the 1970s, and categorize skin color to six tones: four of the skin for “white” skin, with the two remaining for “chocolate” and “black” skin.
Google has confirmed it looks at a different approach, although it is not ready to spill details. “We are working on alternatives, more inclusive, steps that can be useful in developing our products,” the company told Reuters, “and will collaborate with scientific and medical experts, as well as groups working with the color community.” It refuses to provide any specifications.
The FST scale is currently designed initially to classify skin types with however it tends to burn the sun. The grouping is then used as part of ultraviolet radiation treatment. But the scale was then adopted more broadly, because the company and organization were looking for ways to group people’s groups.
The grouping does not have to be as far as engine learning and AI too. Even Emoji looks at the FST for the selection of skin tones, for example.
It is not clear what work may be done by other technology companies on a more effective and balanced scale. Google has provided a timeline when it might have an alternative system ready to be placed; The possibility that, even when it is considered at the appropriate stage for the use faced by the public, the implementation will be carried out calmly on the backend so that it must avoid the hiccups launched from visible. Apple and Microsoft confirmed with Reuters that they realized the shortcomings of FST, even though they were actively developing modern partners for it’s uncertainty.