🚨 Independent AI

Today Korean Social News for Beginners | 2026.01.16

0️⃣ South Korea's Race for AI Technology Independence and National Security

📌 "Chinese Models Risk Security Leaks"…South Korean Government Strengthens Independent AI Standards, Excluding Naver

💬 The government excluded Naver from its "Independent AI Foundation Project" to select Korea's national AI, prioritizing independent technology. The Ministry of Science and ICT determined that using foreign AI models or Chinese open-source technology could create national security risks including information leaks. As a result, LG AI Research, SK Telecom, and Upstage—companies developing independent domestic models—entered full-scale competition. The government stated that securing AI control and sovereignty is a core mission. Selected companies are now competing in industrial applications and technology advancement. Meanwhile, excluded companies like Naver and NC AI plan to respond by developing their own AI chips and expanding services.

💡 Summary

• Independent AI means artificial intelligence developed entirely with domestic technology without foreign dependence.
• The government strengthened independence standards due to national security concerns about Chinese open-source technology.
• LG, SK Telecom, and Upstage were selected, launching Korea's national AI development race.

1️⃣ Definition

Independent AI refers to artificial intelligence models developed entirely from scratch with domestic technology and personnel, without relying on foreign technology or open-source models. It's not just about modifying or training existing models, but independently designing and implementing core algorithms, data processing methods, and learning structures.

The government's emphasis on independent AI centers on technology sovereignty and security. As AI becomes deeply embedded in national infrastructure including defense, telecommunications, finance, and healthcare, dependence on foreign technology could lead to data leaks or loss of technology control. Concerns are growing that using Chinese or American open-source AI could make Korea subject to policy changes or technology restrictions from those countries.

💡 Why This Matters

• As AI is used in national infrastructure, technology sovereignty directly relates to national security.
• Foreign model dependence risks data leaks, loss of technology control, and policy subordination.
• Securing independent AI provides the foundation for long-term industrial competitiveness and technology self-reliance.
• It's essential for establishing Korea as a technology-independent nation in global AI competition.

2️⃣ Current Status and Issues of Independent AI Projects

📕 Government's Independent AI Foundation Project

• A project to select Korea's national AI has begun. Key background:

  • The Ministry of Science and ICT launched the "Independent AI Foundation Project" to select national-level AI models.
  • It determined that using foreign AI models or Chinese open-source technology poses significant security risks including information leaks.
  • Complete technology independence was set as the core criterion to secure AI control and sovereignty.
  • Selected companies receive large-scale government support and priority in national projects.

• LG, SK Telecom, and Upstage were finally selected. Key features:

  • LG AI Research is strengthening industry-specific AI based on its self-developed EXAONE model.
  • SK Telecom is expanding AI services combined with telecommunications infrastructure and strengthening global cooperation.
  • Upstage was recognized for its lightweight, high-efficiency AI model practicality and is expanding corporate services.
  • All three companies share the commonality of developing models with independent technology without relying on foreign open-source.

📕 Background of Naver and NC AI's Exclusion

• Naver was excluded due to lack of independence. Main reasons:

  • Naver's HyperCLOVA X reportedly used some Chinese open-source AI technology.
  • The government saw high risks of information leaks and technology dependence when using Chinese technology.
  • It judged that basing on foreign open-source models makes complete control over core algorithms difficult.
  • The principle was applied that even with excellent performance, not meeting independence standards disqualifies it as national AI.

• Excluded companies are also preparing response strategies. Main directions:

  • Naver plans to strengthen technology independence by expanding investment in developing its own AI semiconductor chips.
  • It aims to maintain commercial competitiveness through personalization services and advanced search and recommendation algorithms.
  • Other excluded companies like NC AI are also considering long-term investments to secure independent technology.
  • They're shifting strategy to secure competitiveness in private markets without government support.

📕 Issues Between Open-Source AI and Independent AI

• Open-source AI has high development efficiency but weak control. Key characteristics:

  • Open-source AI allows anyone to access, modify, and distribute, significantly reducing development time and costs.
  • It enables rapid performance improvement by utilizing global community knowledge and technology.
  • However, making source code public can expose security vulnerabilities and weakens control over core technology.
  • Concerns about technology dependence and security are growing especially with China-led open-source expansion.

• Independent AI has strong control but high development costs. Main pros and cons:

  • Developing all technology in-house ensures complete control and security.
  • It's free from foreign policy changes or technology restrictions, enabling long-term technology independence.
  • However, it requires enormous time, costs, and personnel for development, and initial performance may be lower than open-source.
  • Compatibility issues with global standards may arise, and international cooperation may be limited.

💡 Key Issues of Independent AI

1. Security vs Efficiency: Independent development is safe but time and cost-intensive
2. Technology Level: Initial independent AI may perform worse than open-source-based models
3. Global Isolation: Concerns about limited cooperation due to compatibility issues with international standards
4. Talent Acquisition: Shortage of advanced AI personnel needed for independent development
5. Sustainability: Difficult for private companies to manage alone if government support ends

3️⃣ Strengthening the Independent AI Ecosystem

✅ Government Support and Industry Development

• Long-term, stable government support is needed. Main directions:

  • R&D budgets for independent AI development companies must be significantly expanded.
  • Technology investment should continue from a 5-10 year perspective rather than short-term results.
  • Private company AI investment should be encouraged through tax benefits and deregulation.
  • Markets should be secured by mandating priority use of independent AI in national projects.

• AI semiconductor and infrastructure investment must be strengthened. Key tasks:

  • High-performance semiconductors (GPU, NPU) needed for AI training must be self-developed or stably secured.
  • Large-scale data centers and computing infrastructure must be built to create AI development environments.
  • Cloud-based AI development platforms should be provided so even small businesses can easily develop AI.
  • Complete technology independence must be achieved by combining domestic AI chips with independent AI models.

✅ Talent Development and Technology Cooperation

• AI specialists must be systematically trained. Main measures:

  • AI-specialized departments and research institutes must be expanded in universities and research institutions.
  • Scholarships and research funding should be provided to core AI talent to prevent overseas outflow.
  • Industry-academia cooperation should be strengthened to provide field-centered practical education.
  • Outstanding overseas talent should be actively recruited and their domestic settlement supported.

• Cooperation and standardization among domestic companies should be promoted. Main directions:

  • A cooperation system should be established where independent AI development companies can share technology and data.
  • Compatibility and interoperability between domestic AI models should be increased to expand the ecosystem.
  • Common datasets and evaluation standards should be prepared to accelerate technology development.
  • An open cooperation platform should be created for small businesses and startups to participate.

✅ Practical Application and Global Competitiveness

• Independent AI must be actively applied to industrial sites. Key tasks:

  • Use cases of independent AI should be created in core industries like manufacturing, finance, healthcare, and logistics.
  • The public sector should take the lead in adopting independent AI to induce private sector expansion.
  • AI solutions and consulting should be supported so companies can easily apply them.
  • Successful practical application cases should be promoted domestically and internationally to prove independent AI's value.

• Strategies for entering global markets must be prepared. Main directions:

  • Domestic independent AI technology should be improved to be compatible with global standards.
  • Technology gaps should be reduced through cooperation with overseas companies and research institutions.
  • Competitiveness should be secured by intensively developing strengths in specific fields (e.g., manufacturing automation, language models).
  • Diplomatic and trade channels should be utilized to actively support domestic AI's overseas expansion.

4️⃣ Related Terms

🔎 AI Sovereignty

• AI sovereignty is a nation's ability to independently develop and operate artificial intelligence technology.
  • AI sovereignty means a nation's ability to independently develop and operate artificial intelligence without depending on external technology or platforms. As a core element of technology sovereignty, its importance grows as AI becomes deeply embedded in national infrastructure and industry.
  • Core elements of AI sovereignty include: first, technical capability to independently develop core algorithms and model structures; second, ability to secure and manage data needed for AI training; third, systems to stably supply AI semiconductors and computing infrastructure; fourth, ecosystems to train and maintain AI specialists.
  • The government emphasizes independent AI because as AI is used in national infrastructure like defense, telecommunications, finance, and healthcare, technology dependence directly leads to security risks. Using foreign AI models can cause problems like data leaks, loss of technology control, and policy subordination. Especially as the US and China weaponize AI technology, the need for technology independence has become more urgent. Securing AI sovereignty has become core to national survival strategy beyond simple technology competition.

🔎 Open-Source AI

• Open-source AI is AI models with publicly available source code that anyone can use and modify.
  • Open-source AI refers to AI where source code, training data, and model structures are publicly available for anyone to freely access, modify, and distribute. Representatives include Meta's LLaMA and China's ChatGLM, rapidly developing through global developer communities.
  • Advantages of open-source AI include: first, greatly reducing development time and costs—modifying existing models for specific purposes is much more efficient than starting from scratch; second, utilizing global community collective intelligence enables rapid performance improvement; third, abundant use cases and documentation make learning and application easy.
  • However, there are disadvantages. First, making source code public can expose security vulnerabilities. Second, lacking control over core technology creates risks of subordination to external policies. Third, using China or US-led open-source weakens technology independence. Fourth, commercial use may be restricted by open-source license conditions. The government emphasizes independent AI to avoid this loss of control and security risks.

🔎 AI Foundation Models

• AI foundation models are large-scale general-purpose AI usable for various tasks.
  • AI foundation models are general-purpose artificial intelligence models pre-trained with large-scale data, usable in various fields by fine-tuning for specific tasks. Representatives include language models (GPT, BERT), image models (DALL-E, Stable Diffusion), and multimodal models (GPT-4V).
  • Characteristics of foundation models include: first, possessing general knowledge through training with vast amounts of data; second, adapting to new tasks with little data through transfer learning; third, performing multiple tasks like text generation, translation, summarization, and Q&A with one model; fourth, continuous improvement with ongoing learning.
  • The government's Independent AI Foundation Project aims to develop such general-purpose models with domestic technology. Since foundation models become the basis of AI ecosystems, independent development enables applying AI with proprietary technology across various industries. EXAONE and GPT developed by LG, SK Telecom, and Upstage are representative examples of domestic foundation models, competing with strengths in industry specialization, telecommunications integration, and lightweight design respectively.

🔎 Technology Control

• Technology control is the right to freely use, modify, and distribute core technology.
  • Technology control means having complete ownership and operational rights over specific technology, freely using, modifying, and distributing without external restrictions. In AI, control over algorithms, model structures, training data, and operational infrastructure is core.
  • Importance of technology control includes: first, stable technology use unaffected by external policy changes; second, minimizing leak risks by self-managing security and personal information; third, developing technology aligned with national strategy; fourth, securing long-term technology competitiveness and industry leadership.
  • Using open-source or foreign AI weakens technology control. For example, when the US restricted AI semiconductor exports to China, China's AI development was hit—depending on core technology on outsiders makes one subject to control anytime. Also, using Chinese open-source AI risks subordination to Chinese government policies. The government seeks to secure complete technology control through independent AI to eliminate external dependence risks and protect technology sovereignty.

5️⃣ Frequently Asked Questions (FAQ)

Q: Doesn't independent AI perform worse than open-source AI?

A: There may be performance differences initially, but securing independent technology is more important long-term.

• Independent AI may initially perform worse than open-source-based models because all technology must be developed from scratch. Open-source AI enables rapid performance improvement by utilizing global community collective intelligence and vast training data. In contrast, independent AI may lack initial competitiveness due to limited resources and personnel.
• However, independent AI is more advantageous long-term. First, as proprietary technology, it has complete control and can be continuously improved. Second, customized AI specialized for domestic industries and culture can be developed. Third, it can develop stably free from external policy changes or technology restrictions. Fourth, securing core technology lays the foundation for global competitiveness. This is why the government chose long-term investment over short-term results. The key is steadily improving independent AI performance through sufficient time and support.

Q: Why was Naver excluded and how will it respond?

A: It didn't meet independence standards due to Chinese open-source technology use, but plans to respond by developing its own chips.

• Naver's HyperCLOVA X performs well but reportedly used some Chinese open-source AI technology. The government strictly applied independence standards and excluded Naver, viewing Chinese technology use as high risk for information leaks and technology dependence. The principle was not recognizing anything less than complete independent development as national AI, regardless of performance.
• Naver is preparing response strategies after exclusion. First, expanding investment in self-developed AI semiconductor chips to pursue complete independence from hardware to software. Second, focusing on independent algorithm research to meet independence standards in next projects. Third, maintaining commercial competitiveness by advancing personalization services and search/recommendation algorithms. Fourth, creating success cases in private markets without government support to prove independent technology value. Long-term, it's expected to build a complete independent AI system and try again.

Q: Doesn't independent AI development block global cooperation?

A: Technology independence and international cooperation can coexist—equal cooperation becomes possible after securing sovereignty.

• Independent AI development doesn't mean rejecting international cooperation. Rather, only after securing proprietary technology can equal-footing global cooperation occur. Without core technology, one must accept one-sided subordination or unfavorable conditions, but with independent technology, mutually beneficial cooperation becomes possible.
• Actually, selected companies are also pursuing global cooperation. SK Telecom has partnerships with overseas AI companies, and Upstage is preparing for global market entry. However, they can cooperate without external dependence risks because they're based on independent technology. The government also supports independent AI to be compatible with global standards and aims to secure international competitiveness by leveraging strengths in specific fields. The key is maintaining balance between protecting technology sovereignty while keeping an open cooperative attitude. Independent AI is the foundation for autonomous, equal international cooperation—not isolation.

View Table of Contents

• Go to Guide