AI's Productivity Paradox: Why Companies Lag

Despite the widespread presence of AI, many companies outside the technology sector are not experiencing substantial improvements in productivity, revenue, or profit margins. This phenomenon mirrors the 'productivity paradox' observed during the early days of personal computers.

Scientist and investor Kai-Fu Lee highlights that while individuals have seen significant benefits from AI, large enterprises face obstacles. These include vast amounts of proprietary data that large language models haven't been trained on, leading to ineffective implementations mostly focused on simple applications like customer service.

Lee advocates for an 'AI native' approach, urging companies to fundamentally rethink their operations. He suggests building an ontology—a comprehensive understanding of the company's processes and data—to enable AI to truly grasp corporate context and provide meaningful insights.

The concept of 'Boss AI' or a 'Super Chief of Staff' is proposed to help CEOs gain deeper organizational transparency. This AI would analyze meeting notes, identify key issues, and track strategy execution, moving beyond what human management alone can achieve.

Diffusion of AI technology is rapid, but realizing its full economic potential for corporations requires strategic transformation. Value is expected to accrue most significantly in applications built upon AI platforms, rather than solely in the underlying models or hardware.

Lee notes that while the U.S. leads in enterprise AI adoption and chip development, China excels in open-source models, hardware, robotics, and consumer apps. He believes the ultimate value will be realized through applications that deliver tangible benefits to end-users and businesses.