[AI Revolution](63) Data Labeling Market's Unrivaled No.1... CrowdWorks

Possessing 500,000 Crowdsourcing Workers
Over 180 Technology Patents for Perfect Quality
70% of Top 30 Companies by Market Cap Are CrowdWorks Clients

"Artificial intelligence (AI) should be considered as something nurtured by human hands."

The core of competitive advantage in AI lies in securing high-quality data. Especially for generative AI, since the results are based on data collected and processed by people, high-quality training data is crucial to improve quality. Moreover, depending on what data is used and how it is used, the model's bias or accuracy can be affected, so as the AI market grows, the demand for skilled data labelers inevitably increases.

Data labeling refers to the process of processing data such as images, documents, voice, and videos so that AI can learn from it. For example, in the autonomous driving field, after collecting numerous road photos for AI to learn, the task is to input what is road and what is an obstacle in the photos. AI learns to distinguish roads and obstacles by repeatedly training on labeled data.

In this data labeling market, there is a company growing rapidly with an innovative method called 'crowdsourcing.' That company is CrowdWorks.

When data is collected and processed by internal company personnel, costs and time inevitably increase, but by utilizing CrowdWorks' outsourced data labelers, costs and time can be significantly reduced.

CrowdWorks maintains excellent data quality through over 180 technical patents to overcome the limitations of crowdsourcing, which is often considered difficult to guarantee quality. Minwoo Park, CEO of CrowdWorks, explained, "Thanks to about 500,000 crowdsourcing workers, we have overwhelming data that latecomers cannot catch up with," adding, "We use a four-step personnel verification system to conduct 100% full inspection of data collected and processed by 500,000 people. This results in a 99% guarantee of data quality."

As of the date of this report, the number of registered users on the CrowdWorks website is 533,538. Although the number of active users varies, the company states that all of them can act as labelers at any time.

The services provided by CrowdWorks include ▲ AI construction service 'CrowdWorks' ▲ integrated solution for AI data construction 'WorksStage' ▲ data labeler workforce matching and dispatch service 'Crowd Jobs' ▲ data labeler education service 'Crowd Academy.'

Based on these services, the company received an investment of 200 million KRW from Naver just four months after its founding. Since then, it has collaborated with about 50 teams at Naver for over six years and is currently participating in data construction essential for AI service development, including Naver's large language model (LLM) HyperCLOVA X.

Currently, it provides customized services to over 430 clients including Kakao, Samsung Electronics, KT, and KB Kookmin Bank. Among the top 30 KOSPI IT companies by market capitalization, 70% are CrowdWorks customers. At the end of last month, it also succeeded in entering the KOSDAQ market.

'Crowd Jobs' and 'Crowd Academy' also contribute to improving employment rates. An analysis of income data from students who took labeling training courses last year showed that among those who earned more than 1 million KRW per month, 80% were previously unemployed and have now established themselves as data labelers.

CEO Park said, "We found that freelancers with irregular income and working hours as well as those in geographically disadvantaged areas are now working as data labelers after completing education," adding, "We will provide more information related to data labeling education to help students find employment."

The company is also strengthening its business-to-government (B2G) projects. In April last year, CrowdWorks established a Public Business Division to respond to data demands from institutions, research institutes, and schools.

CrowdWorks' next goal is global expansion. Based on the know-how and technical capabilities accumulated domestically, the company plans to actively enter overseas markets such as Europe, the Americas, and Japan to expand its market.

CEO Park said, "Recently, overseas companies are adopting generative AI models tailored to their businesses to improve work productivity and efficiency," adding, "We will design and build customized data for companies and develop language models that meet their needs to lead the overseas market."

This content was produced with the assistance of AI translation services.