Key Takeaways
- • Motif Technologies disqualified from South Korea's sovereign AI project despite posting the highest score on global benchmarks, marking a major reversal in the national AI competition and sparking controversy over evaluation methodology.
- • Three finalists advance: LG, SKT, and Upstage secured positions in the final phase of South Korea's "National Representative AI" selection after Motif was eliminated in the second evaluation stage, which weighted benchmark scores at 40 points, expert evaluations at 35 points, and other criteria.
- • Benchmark score manipulation allegations plague evaluation: Multiple participants faced accusations of collaborating with foreign firms to artificially inflate benchmark performance, though all four competing teams denied involvement in the controversy.
- • Naver and NC previously eliminated: The first evaluation phase already removed Naver Cloud and NC AI from contention, narrowing an initially broader field of competitors seeking to develop South Korea's proprietary AI model.
- • Strategic vulnerability exposed: South Korea's AI sector faces concurrent brain drain concerns, with the nation losing 1,716 top AI researchers between 2011 and 2024 according to World Bank data, creating talent supply risks alongside the competitive selection process.
Summary
South Korea's government announced the three finalists in its "National Representative AI" selection process on August 27, 2026, but the decision generated significant controversy following the elimination of Motif Technologies—the only team disqualified in the second evaluation stage despite achieving the highest scores on global AI benchmarks.
The evaluation framework employed a multi-criteria approach combining benchmark evaluations (40 points), expert assessments (35 points), and real-use category evaluations, with results varying by target group. The government disclosed detailed scores for all participants, revealing a paradoxical outcome: the team with superior benchmark performance was cut while three other contestants advanced. LG, SKT, and Upstage secured advancement to subsequent phases of the national AI development initiative.
Motif Technologies' elimination sparked outcry within South Korea's AI community, particularly given the team's demonstrated strength on internationally recognized benchmarks. The controversy intensified amid pre-existing allegations that some participating teams had engaged in collaborations with foreign entities to artificially inflate their benchmark scores. All four competing teams publicly denied involvement in any benchmark score manipulation, though the accusations underscored underlying tensions around evaluation methodology and international competitive dynamics.
The sovereign AI selection process represents a critical strategic initiative for South Korea as the nation seeks to develop proprietary artificial intelligence capabilities. However, the evaluation's credibility faces additional pressure from documented talent outflows: World Bank data shows South Korea lost 1,716 top AI researchers during the 2011-2024 period, signaling a broader brain drain that threatens the nation's AI competitiveness independent of domestic selection controversies.
The evaluation framework employed a multi-criteria approach combining benchmark evaluations (40 points), expert assessments (35 points), and real-use category evaluations, with results varying by target group. The government disclosed detailed scores for all participants, revealing a paradoxical outcome: the team with superior benchmark performance was cut while three other contestants advanced. LG, SKT, and Upstage secured advancement to subsequent phases of the national AI development initiative.
Motif Technologies' elimination sparked outcry within South Korea's AI community, particularly given the team's demonstrated strength on internationally recognized benchmarks. The controversy intensified amid pre-existing allegations that some participating teams had engaged in collaborations with foreign entities to artificially inflate their benchmark scores. All four competing teams publicly denied involvement in any benchmark score manipulation, though the accusations underscored underlying tensions around evaluation methodology and international competitive dynamics.
The sovereign AI selection process represents a critical strategic initiative for South Korea as the nation seeks to develop proprietary artificial intelligence capabilities. However, the evaluation's credibility faces additional pressure from documented talent outflows: World Bank data shows South Korea lost 1,716 top AI researchers during the 2011-2024 period, signaling a broader brain drain that threatens the nation's AI competitiveness independent of domestic selection controversies.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.