What has invited the attention of the whole world is the emergence of AI models ChatGPT and its versions. It has led to a need for data annotation or labelling of the data. Data annotation is the process of labelling individual elements of training data to facilitate the understanding of machines about the nature of the data and the significance to be attached to the elements. Such annotated data is used for model training. In other words, we curate the data that goes into the training of models. Of late, data annotation has seen exponential growth. Data annotation will be an industry worth $12 billion by 2030.

Two things are necessary for data annotation — trained manpower and efficient annotation platform.

These are the days of commercialisation of AI. Some annotators are supervised. AI is being applied to different levels of production. There is an explosion of data. There has to be sifting and triaging of data. It is fed into time and test cycles. It involves considerable amount of work. This requires talented manpower trained in data annotation. In addition, there is unsupervised learning in this ChatGPT days. There is also semi-supervised learning. In such situations, there should be people in the loop.

In automobile industry, there is vehicle-to-vehicle communication and connected car technology. There are areas of speech recognition and NLP. Data annotation thus becomes an integral part of this industry.

RWS, UK-based company, has introduced TrainAI. It is an end-to-end data collection, data annotation and data validation service for all types of AI data. It employs a pool of 1 lac annotators who provide AI services in 400 plus languages in 175 countries. They collect, annotate and validate any type of AI data. These annotators must have diverse talents. A part of the work focuses on visuals. The next leg of AI machines will have big focus on visuals. Soma, a US-based non-profit organisation, focuses on visuals.