AI capabilities have emerged beyond the private sector, becoming a modern mission essential for national survival. Consequently, efforts toward AX (AI Transformation) are accelerating within the public sector. Last December, the Korean government announced its plan for the AI transformation of the public sector as one of the "Four Strategies for Government Innovation for a People-Sovereign Government." Innovation in AI-based, task-specific Robotic Process Automation (RPA) was presented as a core task, and in line with this, major ministries and their affiliated public institutions are actively seeking to reshape the paradigm of administrative services through the utilization of AI.
However, AX for public institutions is no simple task. According to a 2025 report published by the NANDA (Networked Agents and Decentralized AI) project under the MIT Media Lab, although global companies have invested over $30 to $40 billion in generative AI technology, only about 5% have successfully achieved meaningful organizational structural changes and tangible improvements in profit and loss through AI integration. The majority remain stalled at the initial Proof of Concept (PoC) or pilot stage, failing to reach real-world implementation.
The report summarizes the common characteristics found in the small number of organizations that have succeeded in incorporating AI: they collaborated with specialized AI solution providers rather than relying solely on in-house development; they set concrete goals led by business departments rather than research institutes or IT centers; and they adopted flexible models that allow AI technology to adapt to business processes in the long term. Reflecting on this, I believe that building symbiotic relationships with AI innovation companies could be the key to success for South Korean public institutions in their AX transition.
By activating collaboration models currently active in the private sector—such as open innovation—tailored to G2B characteristics, we can induce in-depth cooperation between AI specialized firms and public institution practitioners. The government also recognizes this necessity; in its "Republic of Korea AI Action Plan," it emphasizes that the government, as the nation’s largest single economic entity, must become the "first customer" of the AI industry through public testbeds and procurement innovation. This serves the dual purpose of acting as a primer for AI innovation companies while simultaneously advancing the AX of public institutions.
However, there is a prerequisite for public-sector AX collaboration to lead to effective business process changes rather than ending as a "showcase" for optics. Data refinement and infrastructure environments must be established first so that private AI solutions can be seamlessly integrated. Furthermore, concrete goals tailored to the needs of business departments must be set, and performance must be verified through metrics that key stakeholders can objectively agree upon. Above all, a customized procurement and bidding system must be in place to ensure that verified, symbiotic collaboration can be sustained.
The partnership between the U.S. Central Intelligence Agency (CIA) and the big data AI firm Palantir Technologies is perhaps the most representative example of such AX cooperation. The CIA not only provided initial venture capital investment through its related venture capital firm, In-Q-Tel, but has also continued its symbiotic relationship as a core client ever since Palantir's founding. Through this process, the CIA was able to significantly upgrade its national security decision-making system based on AI data analysis, while Palantir was able to establish itself as a leader in the global AI industry using this as a reference.
In this era of technological upheaval, traditional relationships and roles must be able to be flexibly redefined as needed. If public institutions, including the government, have hitherto played the role of providers offering policy support to tech companies, they must now also become active consumers of technology for the sake of AX. By doing so, they can, in the short term, multiply their own AI competitiveness, and in the long term, serve as testbeds and references, thereby fulfilling the core mission of public institutions while fostering the foundation for the growth of technology companies.
