Real estate agents in the United States can no longer imagine working without ChatGPT. “I recently wrote a description for a four-bedroom home using ChatGPT,” said JJ Johannes, an Iowa realtor. He created the description in less than five seconds by typing a few keywords into ChatGPT. “The task would otherwise have taken me over an hour to write on my own,” he told CNN.
Using artificial intelligence (AI) at work is becoming commonplace everywhere, including in South Korea. “When I’m working in the office, I typically search for case laws manually, but when I’m in a hurry or pressed for time, I often rely on generative AI search engines such as Public City,” said Park Si-young, a lawyer at law firm Lin. “Recently, I inquired about random auctions, and the AI promptly delivered relevant precedents, which I found comparable to the research capabilities of a junior lawyer.”
As the number of AI natives - those who skillfully use AI for work - grows, there are concerns that the ‘AI divide,’ a phenomenon more daunting than the digital divide, is becoming a reality. The digital divide is the gap between those who can use digital devices and those who cannot. Similarly, the AI divide refers to the gap between those who can effectively utilize AI and those who are less adept.
“AI is increasingly becoming ‘doping for knowledge workers’ in the same way doping boosts performance for athletes,” said Frederik Anseel, a professor at the University of New South Wales (UNSW) in Australia. AI brings tremendous productivity gains to the workforce, widening the gap between those who can use AI effectively and those who cannot.” The WEEKLY BIZ team at the Chosunilbo recently examined the great AI divide to uncover its implications for the future of work.
AI mastery defines success in modern workplaces
Recently, workplace efficiency has increasingly been determined by one’s proficiency with generative AI. This technology allows some employees to whip up reports or meeting minutes in minutes, while others struggle for hours. Choi, a 42-year-old PR professional at a multinational company, lived in English-speaking countries for 18 years and is fluent in English. However, when creating official English documents, he uses ChatGPT and the AI translation program DeepL to draft the initial version, only focusing on revisions and review. “Using AI to draft the English text reduces a task that used to take two hours to about 30 minutes,” he said.
The boost in work efficiency through AI is supported by statistics. A notable study from September last year, “Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality” by Harvard Business School, highlights this. The report quantified the differences in productivity between 758 consultants at Boston Consulting Group (BCG) who used ChatGPT-4 and those who didn’t. Those who utilized ChatGPT-4 produced 12.2% more work and completed tasks 25.1% faster than their counterparts. Furthermore, when tasked with generating new product ideas, the AI-using group produced 42.5% higher quality results than those who did not use AI.
Another similar survey by Microsoft, released in November last year, gathered feedback from 297 users of their generative AI app, Copilot. According to the survey, 70% of Copilot users reported increased productivity, and 68% noted an improvement in work quality. Frederik Anseel, a professor at the University of New South Wales, commented in his column, ‘The productivity divide: how AI will separate the strong from the weak,’ saying, “AI can better be thought of as having an unlimited army of knowledgeable interns at one’s disposal; an army of interns that are ideally fit for brainstorming, design jobs, writing drafts, analyzing, and coding. Generic AI will program and run other, more specialized AI to do jobs; AI will manage other AI that will manage other AI – the growth and scaling up of this AI productivity chain will be exponential.”
This ‘magical’ increase in productivity may also lead to wage disparities, as predicted by the McKinsey Global Institute. In a report presented at the World Economic Forum, McKinsey Global Institute predicted that by 2030, about 13% of total wages would shift toward jobs requiring high-level digital skills, resulting in wage increases for those roles. Meanwhile, workers with lower digital skills may experience wage stagnation or decline.
64% of Korean workers don’t use AI at work
However, the culture of utilizing AI in the Korean workplace has yet to be widespread. WEEKLY BIZ commissioned SM C&C survey platform Tillian Pro to survey 1,173 office workers in their 20s to 50s on May 10 and 11. The survey found that only 422 (36.0%) had used generative AI, such as ChatGPT, in their work, while 751 workers (64.0%) had never used it. When broken down by age group, workers in their 20s were the most likely to use AI at work, with nearly half (47.6%) reporting usage. In contrast, the usage rates among workers in their 30s (32.4%), 40s (34.3%), and 50s (33.8%) were much lower, hovering around the low 30% range.
In an online survey conducted last year by the Media Research Center of the Korea Press Foundation among 1,000 people in their 20s to 50s, only 32.8% said they used ChatGPT, with paid users accounting for just 5% of total respondents.
Although survey methods may vary, the gap between AI utilization among workers overseas and those in Korea has widened significantly. A survey conducted last year by the global HR services firm Adecco Group among 30,000 employees in 23 countries found that nearly 70% were using generative AI at work.
According to the WEEKLY BIZ survey, Korean workers believe that the performance gap between those proficient in AI and those who are not will grow, but many are not making significant efforts to improve their AI skills. When asked if they agreed that the performance gap between those proficient in AI and those who are not will grow in the future, 71.4% of Korean workers said they agreed. However, when asked about their efforts to utilize AI effectively, aside from “studying how to use it by watching YouTube or searching the internet” (40.3% - multiple responses), responses such as “taking related classes” (17.1%) or “reading related books” (10.7%) were low. Additionally, 33.3% of respondents said they had made no effort.
People born in the 1970s and 1980s may feel intimidated by AI
The digital divide, which is characterized by how proficiently one can use digital devices, often varies significantly by age. Generally, younger people are familiar with digital devices, while older people are less adept.
However, the new trend of the AI divide is emerging even among relatively young generations in their 30s and 50s. The WEEKLY BIZ survey also indicated that office workers in S. Korea, aged from 30s to 50s, use AI at similar rates and primarily for basic tasks like information searches and foreign language translations.
Hong Ki-Hoon, Associate Professor of Hongik University’s College of Business commented, “Unlike the digital divide, the AI divide occurs even among the younger generation. There is a gap between those who have been using AI tools consistently and those who are unfamiliar with them.”
He further noted, “This disparity arises based on the urgency of the need for AI in their work, whether accuracy is prioritized over speed, and whether their workplace encourages AI utilization.”
Consequently, if individuals do not proactively learn new technologies, the gap in AI proficiency among them is likely to widen. Just as the digital divide has left many older people intimidated by ordering from kiosks, many people born in the 1970s and 1980s might feel overwhelmed by AI.
Experts emphasize that the ability to craft effective ‘prompts’, commands given to AI, determines one’s AI utilization skills. While advanced AI utilization might require some training, basic skills can be learned within a few hours. This underscores the urgency of expanding AI re-education programs and offering both online and offline courses.
Anant Agarwal, the founder of edX and a professor at MIT, stated on CNBC, “The more proficient you are with prompts, the more effectively you can perform tasks like writing emails, reports, and creating PowerPoint presentations,” adding, “Everyone can learn the basic prompt skills in about two hours.”
Companies are increasingly embracing AI
The AI gap is evident not only among individuals but also among companies. Some companies are already rapidly evolving by leveraging AI. According to the New York Times, American companies are actively using AI in various fields such as store inventory management and clothing design. For example, ice cream company Ben & Jerry’s has installed AI-enabled cameras to monitor grocery store refrigerators and inform distributors in real-time about product shortages. Catherine Reynolds, a spokesperson for Ben & Jerry’s parent company Unilever, said, “Stores with AI cameras saw a 13% increase in sales because the most popular ice cream products were quickly replenished.”
American clothing company Abercrombie & Fitch enhances work efficiency by using the AI image generation program Midjourney during design idea meetings. Agricultural machinery company John Deere uses AI cameras to more precisely identify areas with weeds, allowing for more efficient herbicide application. John Deere reported that this technology saved 8 million gallons (about 30.28 million liters) of herbicide last year. Department store chain Macy’s employs generative AI for personalized marketing, such as sending emails to customers and adding product descriptions online. As companies that actively adopt AI in marketing and sales achieve higher performance, the performance gap with companies less adept at using AI is inevitably widening.
In a survey conducted by Deloitte Global and Fortune this February, which involved 107 global CEOs, 58% reported adopting generative AI for task automation, and 48% planned to implement generative AI beyond automation. This indicates that AI technology is rapidly being integrated into companies.
MIT Sloan School of Management professor Nathan Wilmers noted in his research report “Generative AI and the Future of Inequality” that it will soon become evident that some companies are much more proficient in utilizing generative AI than others, leading to significant gaps between companies and varying levels of employee wages. The McKinsey Global Institute also predicted at the World Economic Forum that the disparity caused by the AI revolution will first become evident among companies. It suggested that innovative and leading companies adopting AI technology are likely to employ more staff and outperform those reluctant or unable to implement AI.
The AI divide is widening at a regional level and across countries
If the scope broadens, AI widens the gap at the individual or enterprise level or between regions within a country and between countries. According to a recent report by Microsoft Research, “The Emerging AI divide in the U.S.,” the West Coast of the U.S., particularly California, has a high average monthly search rate for ChatGPT, while states like Louisiana, Alabama, and Mississippi have been categorized as having a low rate of ChatGPT searches. Within the U.S., the report revealed that urbanized, higher-income, and more educated regions, along with areas having a larger Asian population and more tech jobs, are more likely to have access to ChatGPT. This disparity contributes to the growing AI imbalance in the country.
Beyond geography, the AI divide between developed and developing countries is already a reality. Developed countries have a high need for AI adoption due to aging populations and high labor costs. In contrast, developing countries are less motivated to adopt AI because of a lack of digital infrastructure and relatively low labor costs for workers. “Many low-income countries do not have the infrastructure or skilled workforce to take advantage of AI, increasing the risk of growing inequality between countries,” International Monetary Fund (IMF) Managing Director Kristalina Georgieva said in an IMF blog. “It is important for countries to build comprehensive social safety nets and provide retraining programs for workers who are vulnerable to AI technologies,” she emphasized.
