As cyber attacks suspected to involve artificial intelligence (AI) agents target the domestic financial sector, funding for national assessments of AI risks has remained stagnant. The budget for the AI Safety Research Institute for the upcoming year is set to increase by 24% compared to this year, but most of this increase is focused on research and development (R&D), leaving actual funding for evaluating AI models and responding to threat trends virtually unchanged.
On October 6, Rep. Lee Jeong-heon of the Democratic Party revealed data from the Ministry of Science and ICT indicating that the government budget proposal for the AI Safety Research Institute in 2027 is 15.944 trillion won, a 23.8% increase from this year.
The overall budget increase is primarily driven by R&D. The budget for R&D focused on developing AI safety technologies for trustworthy AI is set to rise from 3.975 billion won to 6.891 billion won, a 73.3% increase, accounting for about 95% of the total budget increase. The ministry plans to continue supporting existing projects on anti-deepfake detection and resilience against adversarial attacks while launching two new projects aimed at enhancing AI safety evaluation and verification technologies.
In contrast, the non-R&D budget for evaluating currently released AI models, collaborating with domestic and international safety organizations, and analyzing threat trends has only increased from 8.9 billion won to 9.053 billion won, a mere 1.7% rise. While funding for long-term evaluation technology development has expanded, the operational budget needed to meet the immediate demand for model evaluations has effectively remained frozen.
The evaluation burden on the AI Safety Research Institute is growing rapidly. The number of AI models evaluated by the institute increased from six in March to eleven in August, an approximately 83% rise. Evaluating a single agent-type AI model typically takes 7 to 10 days, with 3 to 4 days required for cyber threat assessments. The cost of running a benchmark can reach 50 to 60 million won.
The institute is also relying on external support for high-performance graphics processing units (GPUs). The evaluation department is currently renting two B200 servers and 16 GPUs from a national AI project to assess large-scale and frontier-level AI models. There are no additional GPUs purchased by the department or confirmed for future acquisition. The institute plans to secure additional resources through rentals.
In response, Rep. Lee Jeong-heon criticized the situation, stating, "While long-term R&D budgets are increasing, neglecting the GPU infrastructure of the 'safety evaluation department' that must verify the immediate risks posed by emerging global frontier AI is a serious case of bureaucratic oversight that risks missing a critical window for national cybersecurity."
He added, "As the possibility of AI autonomously exploring vulnerabilities and designing infiltration paths becomes a realistic threat, it is essential to expand dedicated evaluation personnel, operational budgets, and computational resources to preemptively verify and control the attack capabilities of AI models, not just respond after the fact."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.