Summary of Key Points
Artificial Intelligence (AI) is considered a major technological revolution, on par with electricity and the internet. Capital markets are highly optimistic about its potential to boost productivity in the future. However, global economic growth has not accelerated accordingly, and there is still pressure from inflation. The reason lies in the constraints imposed by various factors such as physical limitations (computing power, energy), organizational structures (business processes), human resources (human supervision), and institutional frameworks (legal regulations). When AI evolves from a tool to an “economic agent” that can make transactions and decisions on its own, it challenges traditional economic assumptions about rational behavior and market competition rules, leading to new issues. For AI to truly trigger a growth revolution, it is not enough to simply overcome these constraints; creating new demand (for products, services, and jobs) is crucial.
Why hasn’t AI significantly improved the economy yet? – Constraints Holding Back Progress
You might often hear about how powerful AI is, but why don’t we see significant changes in economic data? This can be explained by the “barrel effect”: the capacity of a barrel is determined by its weakest link. Currently, AI is hindered by several key limitations:
- Physical constraints: AI requires substantial computing power (e.g., GPU chips) and energy (data centers consume a lot of electricity), as well as physical devices (robots, sensors). For example, if a factory wants to use AI to optimize production but doesn’t have enough robots to execute the instructions, even the most intelligent AI will be ineffective. Building data centers also depends on a robust power infrastructure, which is not always available in all regions.
- Organizational constraints: Many companies merely apply AI to existing processes without reengineering them. For instance, although AI can help with report writing, if the approval process remains cumbersome, its efficiency remains unchanged. Incompatibility between old databases also prevents AI from integrating information effectively.
- Human resource constraints: The results generated by AI need to be verified by humans. Lawyers must check contracts for errors, and doctors must confirm diagnoses. Sometimes, AI even increases the workload rather than reducing it.
- Institutional constraints: Laws and regulations restrict the use of AI. For example, handling medical data raises privacy concerns, and companies are hesitant to use AI in this context. There is also a lack of clarity around liability in case of issues, which discourages widespread adoption.
These constraints prevent AI from significantly improving overall economic productivity.
What problems will AI pose to economics when it becomes an “economic agent”?
Previously, AI served as a tool to assist humans; now, it can make decisions on its own (e.g., trading stocks or negotiating contracts), which challenges traditional economic theories:
- Mismatch of goals: AI optimizes according to set objectives, but these may not align with human priorities. For example, delivery platforms might use AI to ensure timely deliveries, leading riders to run red lights, which is efficient but not always safe.
- Algorithms colluding: Multiple AI systems could secretly coordinate to set high prices. This resembles monopolistic behavior, forcing consumers to pay more without clear anti-monopoly laws to regulate it.
- Homogenized decision-making: If many companies use the same AI models, they will make similar decisions, potentially causing market crashes (e.g., if everyone buys or sells stocks based on the same predictions).
- Cultural decline: Relying too much on AI for information could reduce the desire for independent learning. Students and researchers might stop thinking for themselves, leading to a decrease in original knowledge creation.
Efficiency improvement is not enough; creating new demand is essential for a real revolution
Historical technological revolutions (electricity, cars, internet) have not only made existing processes faster but also created new products and services. For AI to trigger a growth revolution, it must do the same:
- Automation as a starting point: While AI can replace jobs (e.g., customer service, report writing), it mainly reduces costs without significantly expanding the economy.
- New demand is the key: AI should create novel products or services. For example, AI-powered doctors could provide remote consultations in remote areas, and companion robots could care for the elderly. These new demands would drive new industries and drive economic growth.
If AI merely optimizes existing processes, its impact on the economy might be less significant than expected. Only by creating new demand can we achieve explosive growth.
Short-term impacts: Possible price increases; long-term potential for explosive growth?
The impact of AI on the economy unfolds in two phases:
- Short term: There may be price increases as companies invest in AI (data centers, chips), leading to higher costs for energy, chips, and metals used in data center construction. This would result in structural inflation (only related to AI-related products).
- Long term: The real potential lies in whether AI can generate new demand. If it creates new markets (e.g., the metaverse, AI-based education, healthcare), economic growth could follow a J-curve pattern.
In summary, while AI is a promising technology, its full impact is yet to be realized due to various constraints. It will also change economic theories and raise new challenges. To truly drive growth, we need to overcome these limitations and create new demand. The road ahead is long, but the potential of AI is enormous.