Zichen Wang delivers a necessary corrective to the breathless hype surrounding artificial intelligence, arguing that the technology's true bottleneck is no longer code, but culture. While most analysts obsess over model parameters, Wang insists we are entering the "second half" of AI development, where the critical challenge is building the invisible social scaffolding required to make these tools actually work. This is a vital perspective for leaders who have invested billions in compute power only to see returns stall in the real world.
The Myth of Technological Determinism
Wang begins by dismantling the most pervasive assumption in the tech sector: the idea that better algorithms automatically lead to better society. "The implicit assumption behind these claims is that technology determines everything: as long as frontier models continue to advance, their social value will inevitably grow in tandem," Wang writes. This framing is sharp because it directly addresses the frustration of executives who have deployed advanced models only to find their workflows unchanged. The author points out that outside a few select sectors, "successful large-scale adoption remains the exception," a reality that contradicts the trillions of dollars in projected growth often cited by industry forecasters.
The piece is particularly effective when it moves from abstract theory to concrete friction. Wang recounts a conversation with the California Public Utilities Commission regarding autonomous trucking. Despite the technology being technically ready for long-haul routes, adoption was stalled not by code, but by "truck drivers' unions strongly opposed its entry." This example perfectly illustrates Wang's core thesis: "There has been much anticipation but little tangible progress" because we are ignoring the human and institutional layers. Critics might argue that union opposition is a temporary barrier that market forces will eventually break, but Wang's historical lens suggests these social resistances are structural features, not bugs, of technological integration.
If the main task in the first half of AI development was technological advancement, then the second half is about its application.
The Automobile as a Mirror
To explain why AI is struggling to land, Wang reaches back to the history of the automobile, a classic case study in sociotechnical systems. He notes that the car did not immediately transform society upon its invention; for the first two decades, it was merely a "status symbol for a small number of wealthy people." It wasn't until the second phase, roughly 1910 to 1945, that a complete ecosystem emerged. Wang details how the technology required a parallel evolution of roads, gas stations, financing, and traffic laws. "The automobile became closely connected to, and interacted deeply with, social culture, industrial networks, national spatial patterns, and the international political landscape," he observes.
This historical parallel is the article's strongest analytical tool. It forces the reader to realize that current debates about data privacy or algorithmic bias are not just regulatory hurdles, but the necessary "traffic lights" and "driver's licenses" of the AI age. Just as the United Nations Committee of Experts on Public Administration has long studied how public institutions adapt to new technologies, Wang argues that AI requires a similar institutional maturation. He writes, "The question facing AI is not only whether the technology itself is mature, but, more importantly, whether a complete supporting sociotechnical system has been established." This reframing shifts the burden of proof from the engineers to the policymakers and organizational leaders.
However, the analogy isn't perfect. Unlike the automobile, which required physical infrastructure like roads that took decades to build, the "infrastructure" for AI is often cognitive and legal, evolving at a speed that physical construction never could. Yet, Wang's point remains: without the social contract, the machine is just a toy.
Building the Intelligent Sociotechnical System
Wang proposes that we must now construct what scholars call an "intelligent sociotechnical system" (iSTS). In this new paradigm, "intelligent machine agents are no longer merely auxiliary tools but collaborative members of human-machine teams." This is a profound shift in how we view the role of AI in the workplace. It suggests that the goal is not to replace humans, but to reorganize organizations so that humans and machines can co-evolve.
He breaks this down into three pillars: physical infrastructure (chips, data centers), institutional governance, and organizational culture. Wang notes that while the U.S. and China are racing to build "intelligent computing centers," the "quality of AI training and inference depends heavily on the quality, representativeness, and reliability of the underlying data." This is where the argument gets practical. It's not enough to have the fastest GPUs; you need the data governance and the cultural trust to use them. "The design objective has likewise shifted from controlling change to responding to it with agility and resilience," Wang asserts. This is a crucial distinction for busy leaders who are used to rigid, top-down management styles.
We are building the engines for a new era, but we haven't paved the roads yet.
Wang also highlights a unique advantage China holds: "China's generally positive public attitude toward AI as an advantage in its adoption and development." This cultural readiness acts as a lubricant for the sociotechnical system, allowing for faster integration than in societies where AI is viewed with deep suspicion. While Western policymakers focus heavily on risk mitigation, Wang suggests that the balance of risk and adoption is a cultural variable that cannot be ignored.
Bottom Line
Zichen Wang's argument is a vital antidote to the technological determinism that dominates current AI discourse, correctly identifying that the next decade of progress depends on social engineering rather than just code. The piece's greatest strength is its historical grounding, which demystifies the current "adoption gap" by comparing it to the slow, messy birth of the automobile age. Its vulnerability lies in the difficulty of executing such a massive sociotechnical overhaul across fragmented global institutions, but the diagnosis is undeniably accurate: we have the engines, but we are still waiting for the roads.