[Policy Pulse]Connected AI, Disconnected Government

원본보기 아이콘

The warning about the future of AI delivered by Eric Schmidt, former CEO of Google, carries significant weight. He points out that if AIs become interconnected and form a collaborative network that creates its own language and communicates independently, it could lead to a situation where humans find it difficult to maintain control. The power of an AI collective that goes beyond individual intelligence and merges together is not an exaggeration but a real existential threat, precisely because its scale and direction are unpredictable.

Yet, here lies a paradox. Unlike AIs that can collaborate to a degree that challenges human control, why is cooperation within human society still so slow to progress? "A single bamboo stalk snaps easily, but a bundle does not break." Humanity has long overcome crises with this wisdom. However, a close look at today's public administration reveals that even this fundamental principle often fails to function properly.

When observing how public officials work, collaboration is emphasized outwardly, but in reality, a fragmented structure focused on ministries, departments, and individuals still dominates. Even when a temporary task force is formed in the name of collaboration, the lead department often sets the agenda while other departments remain at the level of merely providing support for requested work. This falls short of genuine collaboration, where the entire process-from defining policy problems to analyzing causes and developing solutions-is carried out under shared responsibility.

Although collaboration has been presented as a core value in past government innovation initiatives, siloed administration still prevails on the ground. The reasons for this persistent structure are clear: when collaboration fails, responsibility is borne by individuals, but when it succeeds, the results are dispersed throughout the organization. Collaboration often becomes a way to blur responsibility rather than share it. When systems fail to transform culture, these silos become entrenched as routine practices, blocking change.

Ultimately, a powerful catalyst is needed-one capable of redesigning the very way we work. Fortunately, that tool is already within our reach: leveraging AI, at the pinnacle of digital technology, as the collaborative infrastructure for all administrative processes. Recently developed AI collaboration tools can instantly connect vast volumes of documents, data, and work context within organizations, enabling many members to participate in analysis and planning simultaneously.

In fact, this transformation has already begun. The United Kingdom is building a collaborative infrastructure that links policy information between ministries through shared AI systems such as Humphrey. The United States is advancing decision-making frameworks that integrate and analyze information from various agencies by utilizing data integration-based AI platforms like Palantir. AI that integrates and analyzes fragmented information is beginning to function as an invisible coordinator, reducing administrative waste and preventing policy conflicts.

As Eric Schmidt's warning suggests, interconnected AI possesses immense power. The issue is not the power itself, but how it is used. Will we leave that power as something to be feared, or will we use it as a catalyst to break down silos within government organizations? When AI is viewed not simply as an automation tool but as an intelligent infrastructure for designing inter-organizational collaboration, its potential becomes reality. The choice belongs to policymakers.

Oh Chulho, Professor Emeritus, Department of Public Administration, Soongsil University

© The Asia Business Daily(www.asiae.co.kr). All rights reserved.