[New Jobs] AI Teaching 'N-jabreo'... Should I Try It Too?

AI Supervised Learning Part-Time Job 'Data Labeling'
Easy Task Support on Platforms
Work Like Playing Games on Smartphones

Editor's NoteWith super-aging and ultra-low birth rates, along with the advent of the artificial intelligence (AI) era, new changes are occurring in the world of jobs. We introduce the world of new jobs that open the path to new challenges in working life or a second act in life.

Training artificial intelligence (AI) requires a vast amount of data. This has become common knowledge, but few people know 'how' to prepare the data to teach AI.

You cannot just feed any data to AI. To create a smart AI, the data must go through a separate verification process. This verification process is called 'data labeling,' and it is all done by humans.

Selecting Data Like a Game to Guide AI Learning

An example of data labeling work. Creating rectangular boxes around objects that AI needs to identify enhances its recognition ability.
[Image source=Testworks homepage]

Specifically, imagine developing an AI that can recognize human hand movements. You need images of human hands to train the AI. However, you cannot randomly gather all hand images and train the AI with them. You must select photos where the palm and fingers are clearly visible. This process of selecting useful data from a massive random dataset to 'supervise learning' for AI is called data labeling.

Although AI is a cutting-edge industry, the actual task of composing datasets still relies on simple human labor. At one time, data labeling was called the 'doll eyeball sticking part-time job' in the IT industry. This means it is a simple and easy task but also quite labor-intensive.

However, it is not all disadvantages. Because it is a simple task, anyone can become a data labeler in a part-time format without special certifications. Most work is done remotely. Above all, labeling part-time jobs are often 'gamified.' A typical example is turning the labeling process into a simple quiz game format. In short, it is an excellent side job where you 'play with AI and earn money' at the same time.

The 'Doll Eye Sticking Part-Time Job' of Advanced Industries... Attractive to N-job Seekers

There is also a platform that allows you to work using a smartphone. [Image source=Getty Images Bank]

Along with the emergence of the AI industry, the data labeling business has also grown. Overseas, it has already established itself as a respectable side job and remote part-time work. As AI grows, the demand for data also skyrockets, so the workload for data labelers increases. The Export-Import Bank of Korea predicted that the global data labeling market size would grow from 10 trillion won in 2021 to around 39 trillion won next year.

There are various data labeling platforms in Korea as well. 'LabelOn,' 'CrowdWorks,' and 'CashMission' are representative examples. These platforms often provide simple in-house training program videos, and recently, many offer work platforms accessible not only via PC or laptop but also smartphones.

However, because of its high accessibility and relatively simple tasks, it is difficult to utilize data labeling as more than a side job. Most data labeling side jobs pay a fixed amount per task, limiting income generation. Nevertheless, for aspiring 'N-job' workers who want to spend their spare time productively and enjoyably, it can be an attractive option.