AI
Economic Paradox with ChatGPT
In recent years, the development of artificial intelligence (AI) has led to a new technological wave that revolutionizes the way people live and work. One of the most significant milestones of AI is the emergence of chatbots – language models that use machine learning to generate interactive text for human interaction. One notable example is ChatGPT, developed by the startup company OpenAI, which has attracted a large user base.
Since its launch in November 2022, ChatGPT has quickly gained global popularity, with over a million sign-ups within just one week, and by the end of January 2023, it surpassed 100 million users. In comparison, according to Sensor Tower statistics, it took Facebook nearly two years, the short video sharing platform TikTok nine months, Instagram 2.5 years, and Google Translate more than six years to reach the 100 million user mark.
Similar to the development of artificial intelligence (AI) itself, ChatGPT has brought about a technological revolution that impacts and transforms the way people live and work. The greatest impact of artificial intelligence will be in enabling humans to achieve discoveries and advancements that would be difficult to attain on their own. ChatGPT utilizes information that humans “have” to generate information that humans would find challenging to obtain, particularly in the realm of prediction. From an economic perspective, what is the impact of ChatGPT?
Does ChatGPT bring about a “revolutionization of the traditional economy”?
In various fields of work, ChatGPT can help expand the expertise and capabilities of non-experts, such as economists with limited knowledge of complex machine learning tools, by providing them with a document search tool. This means that machine learning can be brought to the field of economics.
To determine the specific contribution of AI to GDP, several factors need to be considered, including the pace of AI technology development, investment and application of this technology across different industries, the development of AI education and training systems, and supportive government policies.
Although there is still limited research on the economic contribution of AI, a report by Goldman Sachs published in April 2023 indicates that AI could increase global GDP by an additional 7% in the next decade. Despite some hesitation and uncertainty from businesses, AI technology is predicted to generate nearly $16 trillion for the global economy by 2030.
That is why many countries around the world are increasing their spending on this technology. Some economists also argue that with the application and development of AI policies, Southeast Asian countries can reap billions of dollars in benefits from artificial intelligence.
In Southeast Asia, AI technology is predicted to contribute $1 trillion to the GDP of countries in the region by 2030. Specifically, AI is expected to contribute $92 billion to the GDP of the Philippines, and for Indonesia, AI could add $366 billion to its GDP in the next decade.
However, behind those monetary figures, artificial intelligence, specifically ChatGPT, is generating benefits that traditional economics has not accounted for, similar to services like free information and data from Wikipedia, email services like Gmail, digital maps like Google Maps, and video-sharing platforms like TikTok. These products have enormous economic value and provide significant benefits, yet they are fundamentally not captured in national accounting systems.
There is a paradox where ChatGPT is creating numerous high-value services but almost has no cost, especially with near-zero marginal costs. In the digital economy, data is increasingly becoming a crucial factor of production. However, accurately measuring the value of intangible assets is challenging beyond establishing their existence.
Therefore, a revolutionary challenge for current traditional economic models is how to integrate the “data” factor with traditional elements, especially human capital, to assess the complete and accurate contributions and sources of growth from each factor.
Can ChatGPT solve the “Solow productivity paradox”?
Paul Krugman has stated that productivity is not everything, but in the long run, it is almost everything when it comes to improving a nation’s standard of living over time. While productivity growth is considered the decisive factor for countries to become wealthier and more prosperous, unfortunately, this process has stagnated in the United States and most advanced economies (specifically, the United Kingdom and Japan) since the 1980s.
In 1987, Robert Solow, who won the Nobel Prize that year for explaining how innovation drives economic growth, famously said, “You can see the computer age everywhere but in the productivity statistics.” This “Solow productivity paradox” refers to the phenomenon in the 1980s and 1990s where, despite rapid advancements in information technology, productivity growth in the United States was negligible.
The discussions and debates among economists about this paradox have raised concerns about whether it applies to AI as well, indicating that AI developers may have somewhat optimistic views and assessments of artificial intelligence technology. Two years ago, there were predictions of a productivity boom from AI and other digital technologies, and today there is continued optimism about the impact of new AI models. Much of this optimism comes from the belief that businesses can greatly benefit from using AI like ChatGPT to expand services and improve labor productivity.
With economies rapidly advancing in AI, the majority of the workforce will consist of knowledge workers and information workers. The question arises as to when productivity growth will fundamentally improve. The answer, in principle, depends on whether we can find ways to use the latest technology differently from previous business transformations in the computer era.
Most economists today believe that patience is required as current statistical data still shows a surprisingly slow impact of artificial intelligence and other digital technologies in improving economic growth. The reason is that so far, companies have only been using AI to perform tasks “slightly better.”
However, with the hope that ChatGPT and other AI chatbots primarily automate cognitive work, which is different from investment in equipment and infrastructure, productivity gains are expected to occur much faster than the previous information technology revolution. Although the timing of productivity growth with AI is still uncertain, some economists have predicted that we may see significant increases in productivity by the end of 2024