이 누리집은 대한민국 공식 전자정부 누리집입니다.

한상넷 로고한상넷

전체검색영역
Korea’s AI technology is at 47% of U.S.’s level: Survey
Collected
2024.02.20
Distributed
2024.02.21
Source
Go Direct
이미지 확대
South Korea’s artificial intelligence (AI) technology stands at only 47 percent of the United States’ level, according to an assessment by local tech developers.

According to a survey conducted by Maeil Business Newspaper on 105 respondents from 59 AI companies and educational institutions about Korea’s AI technology status on Sunday, Korea’s AI technology level averaged 4.7 points on a scale of 1 to 10, with the U.S. as the benchmark.

The survey comes amid rapid expansion of the AI ecosystem among major global technology firms. The evident trend in the global AI industry this year is the rise of AI agents, capable of understanding human speech and executing various tasks for users.

Industry insiders predict a three-way competition between OpenAI’s ChatGPT, Microsoft Corp.’s Copilot, and Google LLC’s Gemini.

To bolster Korea’s AI competitiveness, 46.7 percent of the respondents pointed to the expansion in AI-related budgets and R&D support, followed by the nurturing of AI talent and recruitment support at 22.9 percent.

Insiders emphasize the need for government-backed infrastructure development to safeguard the domestic market, given the advance of big tech giants leveraging their technological prowess and financial resources.

According to Goldman Sachs Group Inc., U.S. big tech firms plan to invest $68.1 billion in AI this year.

Expansion of graphic processing unit (GPU) infrastructure is a critical task. Despite being recognized as vital infrastructure, GPU supply has consistently fallen short of demand. However, this development presents a significant cost challenge for domestic companies, requiring billions of dollars.

An unnamed researcher suggested consolidating government budgets toward GPU infrastructure investment, rather than allocating them for various purposes.

Active data acquisition was deemed crucial for GPU development among the respondents as well. “Securing GPUs demands significant capital, but data acquisition requires relatively smaller costs,” said Kim Nam-hyuk from AI company Upstage.

By Won Ho-sup, Lee Sang-duk, and Chang Iou-chung

[ⓒ Pulse by Maeil Business News Korea & mk.co.kr, All rights reserved]