Commentary: Beijing’s Plan to Unlock AI’s Economic Potential

The global investment in artificial intelligence (AI) IT reached $315.8 billion in 2024 and is projected to grow at a compound annual rate of 32.9% through 2028. Despite this large influx of capital, fewer than 5% of current AI projects are generating significant productivity or profit gains, as found in an MIT survey, highlighting a lag between investment and macroeconomic payoff. The data show that the AI revolution is ongoing but that its major economic benefits are still accumulating rather than fully realized at this stage. [para. 1]

Against this backdrop, China’s State Council released the “Opinions on Deeply Implementing the ‘AI+’ Action,” which charts a comprehensive strategy for integrating AI across all sectors. This policy initiative emphasizes leveraging AI to transform both economy and society, fostering revolutionary changes in productivity and production relations. The policy aims to channel AI’s potential into tangible economic dividends through deliberate and top-level planning; under this guidance, the economic impact of AI is expected to accelerate and become a defining feature of a new era of productivity. [para. 2]

Currently, AI’s main value lies in reducing costs and boosting efficiency, especially in tasks like financial settlement and compliance. Some firms have reported annual savings of $2–10 million due to AI-driven back-office automation. However, these short-term gains are insufficient for a macro-level leap in productivity. The long-term potential of AI will depend on bridging the gap between technological breakthroughs and large-scale industry adoption. Most industries remain in the pilot phase, with only tech and media showing meaningful initial transformation. AI’s macroeconomic impact is predicted to follow an “S-curve” model: initial high investment brings limited growth, which rises as applications broaden and then ultimately triggers major productivity gains through cross-industry integration. Policy leadership remains crucial in this development phase, as outlined in the State Council’s Opinions, which call for policy-guided breakthroughs to accelerate AI integration across primary industries and public services. [para. 3][para. 4][para. 5]

A structural mismatch currently exists in global AI investment. Around 50-70% of budgets are funneled toward sales and marketing, seeking quick returns, while back-office areas—where AI’s impact can offer higher returns—receive less funding. This imbalance results in reduced capital efficiency, widens gaps between industries (with manufacturing and transportation lagging), and underscores the importance of policy in guiding investment towards sectors with high productivity potential, like manufacturing, agriculture, and healthcare. The Opinions address this by advocating for the use of industrial funds and incentives to encourage AI adoption in under-invested sectors, promoting a comprehensive and balanced productivity boost. [para. 6][para. 7][para. 8][para. 9][para. 10]

On the labor market, fears of imminent mass unemployment from AI automation remain overstated. AI’s substitution effect is predominantly limited to non-core departments, with layoff rates in the range of 5–20%. A more significant shift is occurring in employment models: AI proficiency is increasingly required, and internal automation is replacing outsourced roles, especially in process-heavy industries. This transition elevates demand for AI-skilled talent and challenges education and training systems to adapt. The Opinions propose incorporating AI into education and promoting new job creation and skills upgrading, aiming to make AI an employment enhancer rather than a threat. [para. 11][para. 12][para. 13][para. 14][para. 15]

In summary, the path to macroeconomic benefits from “AI+” will be prolonged but promising, relying on patient accumulation of capabilities and strategic policy support. With clear institutional direction and market participation, China aims to drive a new era of smart economic development, mirroring the transformational roles of electricity and the internet in previous eras. [para. 16]