The provided report outlines a controversy regarding the transparency and administrative conduct of the South Korean Ministry of Science and ICT (MSIT) concerning the "Autonomous AI Foundation Model" (DocPaMo) project, a national initiative with a 530 billion KRW budget running through 2027.

Core Conflict and Administrative Action

Motif Technology, a participant in the project's second stage of evaluation, was eliminated in August 2026. Despite having a model that achieved an AAII (Artificial Intelligence Index) score of 47—ranking 1st outside of the U.S. and China and outperforming the selected consortia (Upstage: 37, SK Telecom: 35, LG AI Research: 31)—Motif was disqualified. In response, Motif submitted a six-page formal objection to the National IT Industry Promotion Agency (NIPA). NIPA processed the objection but issued a single-word notification: "Dismissed" (기각). This rejection lacked any substantiation, explanation of the evaluation criteria, or responses to the specific queries raised by Motif, prompting criticism regarding the fairness and transparency of the government's procurement process.

Evaluation Methodology and Technical Grievances

The DocPaMo second-stage evaluation was conducted on a 100-point scale based on the following weighted distribution:
$$E_{total} = (B \times 0.40) + (E_p \times 0.35) + (U \times 0.25)$$
Where:

• $B$ represents the benchmark score (e.g., AAII).
• $E_p$ represents expert qualitative evaluation.
• $U$ represents user experience/evaluation metrics.

Motif’s primary technical grievances focus on the lack of clarity regarding the following:

1. Benchmark Normalization: The methodology for calculating the distribution of the 40-point benchmark score given the discrepancy between Motif's high AAII score and their overall elimination.
2. Qualitative Criteria: The specific metrics and decision-making logic used by expert panels ($E_p$). Motif noted that evaluators questioned the company regarding their "consortium lead" status and the involvement of foreign capital or overseas corporate entities as major shareholders with voting rights. Motif requested disclosure on whether these specific queries influenced the scoring.
3. Transparency of Process: The absence of a detailed breakdown of scores across the three categories ($B, E_p, U$), which prevents the company from understanding the causal factors behind their disqualification.

Conclusion and Future Outlook

Motif asserts that their objection was not aimed at overturning the selection, but rather at fostering an industry-wide consensus on the project's objectives and establishing a constructive feedback loop for future developments. Following this incident, Motif has declared its intent to abstain from the third stage of the project. The MSIT has stated that working-level staff will meet with Motif to provide further explanations; however, the incident highlights significant concerns among stakeholders regarding the administrative accountability of large-scale national AI R&D funding.