Balancing Innovation and Competition in the Age of AI

In the 1980s, Peter Howitt and I developed the Schumpeterian growth paradigm to analyze how this process fosters economic growth. The model suggests that long-term growth arises from a continuous cycle of innovation, where each new innovator builds upon previous advancements. Entrepreneurs engage in research and development as successful innovation can yield temporary economic benefits, but these benefits can also be used to protect established positions and stifle further innovation.

Competition plays a critical role here. Companies that are near the technological frontier tend to innovate more in response to increased competition, while those farther behind may retreat, hindered by stronger market forces. Thus, a dynamic economy needs to maintain both incentives for innovation and pathways for new challengers.

The information technology revolution in the United States illustrates this challenge well. From the mid-1990s to the early 2000s, the IT sector saw significant productivity growth, driven by leading companies like Google, Microsoft, Amazon, and Walmart. However, as these companies became more dominant, they discouraged market entry by new firms, demonstrating how yesterday's innovators can obstruct future ones.

As economies transition from adopting to creating frontier technologies, challenges can arise. South Korea and Japan achieved significant growth by imitating advanced technologies, but there comes a point when merely imitating is insufficient for sustained growth. Established corporate groups may resist competition and inhibit new entrants, as evidenced by South Korea's chaebols and Japan's keiretsu.

In contrast, China is gaining ground in frontier innovation, increasingly contributing its own innovations rather than relying solely on foreign technologies. This makes the discussion about artificial intelligence (AI) particularly relevant.

While some view the AI revolution with skepticism, I contend that it does not need to be feared if supported by appropriate institutions. AI has the potential to significantly enhance productivity by automating various tasks, both in goods production and idea generation. Our estimates suggest that AI-driven automation could boost productivity growth by about 0.68 percentage points annually over ten years, with potential increased gains from AI's impact on new idea creation.

However, AI's rise also mirrors competition issues seen previously. The AI value chain is becoming concentrated, dominated by major tech companies in cloud computing and graphics processing. This indicates that despite its growth potential, technology can fail to deliver sustained benefits if competition declines.

Employment presents another paradox. While AI may displace certain tasks, its overall impact on employment need not be negative. Companies utilizing AI can improve competitiveness, leading to increased demand for their products and subsequent hiring. Moreover, AI can hasten the development of new ideas, generating economic activity and jobs.

The pressing question, therefore, is how societies will harness AI. To realize its growth potential, we require effective competition policies that enable new ideas to challenge established firms. Additionally, strong educational and labor market systems are necessary to assist individuals in transitioning into new roles through retraining.

Ultimately, the lesson of creative destruction is not that disruption is inherently negative or positive, but rather that economies must facilitate successful innovation while preventing past winners from obstructing future progress. With the right institutions in place, creative destruction can foster both rapid and inclusive growth, capturing the promise of the AI revolution.