Institute to Transform Science Research with AI Integration
South Korea will launch the National Science AI Research Institute in the first half of the year, with an opening targeted for June. The move was outlined during work reports delivered to the Ministry of Science and ICT on January 12, led by Deputy Prime Minister and Minister of Science and ICT Bae Gyeong-hun. It signals a clear shift: government-funded research institutes will start retooling their research structure around AI.
What's launching and who's involved
On January 12, 21 government research institutes and 28 public institutions presented their research and management plans. The National Research Council of Science and Technology (NST) detailed the plan for the National Science AI Research Institute as a central hub linking science, technology, and AI across institutes.
KIST has already secured 40 billion won to build AI infrastructure, operational systems, and support functions for the Institute. The aim is to open in June, with initial services focused on government research institutes.
How the AI "co-researcher" model will work
The Institute will pilot an AI co-researcher model across government labs. In practice, AI will partner with scientists on experimental design, data collection and analysis, hypothesis testing, and related tasks. The goal is straightforward: raise throughput and shorten time-to-results without lowering scientific rigor.
Research planning and operations are being rewired
With the project-based system (PBS) abolished, planning and coordination are being rebuilt. KIST will set up a dedicated unit in February to support strategic research projects and strengthen cross-group planning around priority technologies.
Evaluation is changing
Starting this month, institutes move to a one-year integrated evaluation that covers both research and management. This replaces the previous cycles of three-year management evaluations and six-year research evaluations. The intent is to reduce short-term, output-only incentives and tighten the link between research strategy and organizational operations.
Timeline and next steps from the Ministry
The Ministry of Science and ICT will continue receiving reports through January 14 from science museums, Korea Post headquarters, postal service public institutions, the Korea AeroSpace Administration, and the four major institutes of science and technology. Expect AI-based innovation strategies to be specified across research and public institutions as these sessions conclude. For official updates, see the Ministry of Science and ICT.
What this means for lab leaders and PIs
If you run a lab or a research program, you don't need a grand overhaul to get value. Start with focused changes that align with the new Institute and the integrated evaluation.
- Identify 2-3 workflows where AI can serve as a co-researcher: design of experiments, data wrangling, simulation pipelines, or routine analysis.
- Make datasets "AI-ready": standardize metadata, institute data versioning, and clarify consent/licensing for reuse.
- Set up validation gates and audit trails for AI outputs to protect reproducibility and compliance.
- Budget for compute and MLOps. Map which needs can be met via shared Institute resources versus in-house capacity.
- Upskill teams on prompting, model limits, bias, and security. A practical place to start: AI courses by job role.
- Align proposals and KPIs with the one-year integrated evaluation, emphasizing strategic outcomes and cross-team coordination over short-term counts.
Why this matters
AI won't replace the fundamentals of good science-clear hypotheses, high-quality data, rigorous methods-but it will pressure-test how efficiently labs operate. The Institute's co-researcher model, combined with reorganized planning and evaluation, gives government labs a common playbook. Teams that prepare their data, processes, and people now will move faster once the Institute opens.
Your membership also unlocks:
