In an era where the responsible deployment of artificial intelligence is paramount, a recent McKinsey report highlights a critical gap: many organizations are ill-prepared to ensure the safety and accountability of generative AI. A significant concern revolves around the explainability of AI's decision-making processes, a challenge recognized by a substantial portion of surveyed businesses, yet actively addressed by only a minority. Against this backdrop, Datumo, a Seoul-based innovator, is stepping forward to tackle this crucial issue with a fresh approach to AI validation and oversight.
Datumo, initially established as a company specializing in AI data labeling, is now pivoting its core focus to empower businesses in constructing more secure and reliable AI systems. This transition is underpinned by their recent successful funding round, which saw the company secure an impressive $15.5 million, elevating their total capital raised to approximately $28 million. This funding influx, drawing support from key investors such as Salesforce Ventures, KB Investment, ACVC Partners, and SBI Investment, underscores the market's confidence in Datumo's vision. The company’s journey began with its founder, David Kim, a former AI researcher, who identified the inefficiencies in data labeling and ingeniously devised a reward-based application, enabling a broader participation in data annotation. This early innovative spirit quickly translated into tangible success, as Datumo, then known as SelectStar, garnered significant pre-contract sales and rapidly surpassed $1 million in revenue within its inaugural year, attracting major Korean clients like Samsung and Hyundai. Over time, client demands evolved beyond mere data labeling, prompting Datumo to recognize its unwitting foray into AI model evaluation, a domain they have since embraced wholeheartedly, culminating in the release of Korea's pioneering benchmark dataset for AI trust and safety.
As the AI landscape continues to evolve rapidly, marked by significant investments like Meta's acquisition-like venture into Scale AI, the need for robust AI evaluation platforms is more pressing than ever. Datumo distinguishes itself in this competitive arena by offering not only pretraining dataset provisioning, akin to industry leaders, but also a unique full-stack evaluation platform, Datumo Eval. This platform stands out for its capacity to automatically generate test data and perform evaluations, identifying potential biases, errors, or unsafe responses without the need for intricate manual scripting. This no-code solution is specifically designed to democratize AI safety, making it accessible to non-technical teams involved in policy, compliance, and trust and safety. The company's strategic growth is further bolstered by this new funding, which will accelerate its research and development in automated enterprise AI evaluation tools and facilitate global market expansion across South Korea, Japan, and the United States, including a newly established Silicon Valley presence, solidifying its position as a significant contender in the global AI evaluation market.
The proactive steps taken by companies like Datumo in fostering safer and more transparent AI systems are vital for the responsible advancement of artificial intelligence. Their commitment to developing accessible and effective evaluation tools reflects a broader societal imperative to ensure that as AI becomes increasingly integrated into daily life, it does so with integrity and trustworthiness, ultimately paving the way for a more equitable and beneficial technological future.
