Eric Brignolpson, director of the Digital Economy Research Institute at Stanford University in the middle of Silicon Valley, emphasized three things throughout his interview with Maeil Economy. The economic value created by artificial intelligence (AI) technology is not properly measured, but it far exceeds our expectations, and the timing of the productivity explosion from AI depends on how quickly the reorganization proceeds through people and work relocation. Finally, AI should be developed and utilized in the direction of expanding capabilities, not replacing humans.

"Consumer surplus generated by Generative AI chatbots in the U.S. is estimated to reach $172.3 billion (about 250 trillion won) this year, which is a nearly 50% increase from $116.2 billion (about 157 trillion won) last year," Brignolpson said. "This is far ahead of companies' AI sales in the U.S." "Consumers enjoy most of the AI benefits at low cost, but it is not reflected in the actual GDP," he explained.

"Consumer willingness to pay and accept digital and AI goods should be measured, and production and sales should be investigated with a new concept of 'GDP-B'," he argued.



There is a lot of debate in economics around the world over the economic effects of AI. MIT Professor Darren Assemoglu (Daron Azemoelu), a Nobel laureate in economics, is at the opposite point. Brignolpson believes that the labor productivity of the U.S. economy will grow twice the past average due to AI, but Professor Assummoglu said he would only increase total factor productivity by 0.5% over the next 10 years.

However, Director Brignolpson said, "I recently spoke with Professor Aschimoglu, and he revised some of his views to reflect the development of agent-type AI."

As for Tesla CEO Elon Musk, who makes the most radical argument, he said, "We underestimated the bottlenecks that appear in every process (introducing AI technology), and technical capabilities do not mean value. Even if AI is introduced, it takes years to spread beyond the barriers of organization, contract, regulation, and trust." It means so-called "J Curve," and it is argued that such organizational redesign can lead to productivity innovation.

"(AI productivity effect) is not a matter of prediction, but a matter of design," he said. "It depends on the choices we haven't made yet, and there is an opportunity to create tremendous abundance using this technology, but there is also a risk of missing that opportunity."

"For companies carrying out reorganization, the depth of the low point (productivity J-curve) will be shallow and the period will be shortened from 15 to 3 to 5 years," Brignolpson said. However, he warned, "If there are no employees who simply insert AI into existing work, consider introduction itself, not performance, or voluntarily document tacit knowledge that can automate jobs, they will quietly fail."

He argued that AI should not replace humans, but create a system that can complement human capabilities. It is the so-called Centaur Evaluation. It is a concept that suggests the direction of coexistence by combining human intuition and creativity with AI capabilities, just like a creature in the myth of anti-human anti-human. Director Brignolpson explained, "AI policies should also be designed to provide incentives to technologies that complement human capabilities rather than technologies that replace humans."

Apart from AI innovation, he advised that new employees should be continuously selected to build skilled personnel. According to Brignolffson's paper "Canaries in Coal Mine, AI's Impact on Employment" published in November last year, AI can kill more than 16% of young workers aged 22 to 25, while it can promote employment by 9% for skilled workers aged 35 and over. "I'd like to add a warning to the management," he said. "Stopping hiring new employees will eliminate experienced personnel in the future in advance, which will be paid when skilled talent is needed in a few years."

Brignolffson was concerned about the rigidity of the Korean labor market. He pointed out that the J-curve recession in Korea could be prolonged, saying, "Organizational redesign is a cost in the productivity J-curve, and labor rigidity increases the burden by blocking the relocation of manpower and work."

[Reporter JINYOUNG TAE]