Five separate sources claim that Google is about to reduce relations with Scale AI. An AI data labeling company that traded with search engine giant.
According to Reuters report, in fact, Google (was?) Scale is the largest user of AI. But what is exactly what AI data labeling firm, of course?
An AI data label company helps employees teach AI by watching a lot of information-such as photos, videos, or text-and adds helpful notes or tags for the explanation of this information. These notes are called “label”, and they tell a given AI model what he is watching or reading. For example, one can see a picture and mark that it contains a dog, or read a phrase and label it with a happy or sad label.
The AI then uses all the labeled data to recognize things, understand the language, or learn how to make a better choice when you ask him to do something or provide information for yourself. Basically, a student needs a teacher to explain things and need such labeled data to learn and improve AI. Without it, AI will not know what to do with all the information he has given.
Picture by Scale AI
The reason for Google’s alleged Scale AI is that Zuckerberg’s Meta (Facebook, Instagram and WhatsApp company) is buying AI on a scale of 49 %. Google was allegedly planning to pay about $ 200 million to use Scale AI services to further improve Gemini in 2025, but now, Big G will probably have to find another player in the city.
It’s not just Google that is turning a step back from Scale AI, apparently: Microsoft, Zee (Elon Musk’s AI project) and 49 % of the Openi Meta’s data labeling company is somewhat nervous about the acquisition of the company. All of them are concerned that if the Meta is partially owned by AI, Meta may have access to its sensitive AI data and research theories.
Of course, that means that a lot of money and business can be lost on the scale, but I personally suspect that they will face any deadly turmoil in the years to come.
At the same time, the scale rivals (such as label boxes, hands shakes, and others) are flooding new users. Some AI Labs are also starting to hire their data labors so they can keep everything private.
Read the latest from Sebastian Pierre
