Article body
The controversy surrounding the 'DOKPAMO' project, which aims to select Korea's representative AI, is intensifying. This project is a key initiative for national AI competitiveness, involving government funding and thousands of the latest NVIDIA chips, and the results of the recent second evaluation were unexpected. While the startup Motif, which scored the highest in the global AI performance evaluation index AAII, was disqualified, a large telecommunications company and a corporate AI research lab with relatively lower scores received passing grades.
The government acknowledged Motif's technological capabilities but faced criticism over the evaluation method. Although there was a significant gap in actual AI performance between the top and bottom contenders, the scoring sheet reduced this gap to just four points. Expert and user evaluations accounted for 60% of the total score, which included not only AI technology but also its applicability in the industry and contributions to the AI ecosystem. This evaluation structure has been criticized for potentially favoring large corporations.
Motif has raised concerns about the fairness of the evaluation method and has expressed its opposition, while the government has left open the possibility of revising the evaluation approach. The essence of the DOKPAMO project is not only about industrial applicability but also about the competition to see who can create the smartest and most autonomous national AI. Therefore, the government needs to provide a more transparent explanation of the validity of the evaluation criteria.