IP system in Korea
Introduction
Patent & Utility Model
Layout Designs
Artificial Intelligence
trademark
design
Key Features
Official Fees

Protection of Artificial Intelligence

Artificial Intelligence (AI) has become an essential part of modern technological advancement, significantly impacting various industries. South Korea, a leader in technology and innovation, has developed a comprehensive legal framework to protect AI technologies. This guide provides an in-depth overview of the mechanisms available for AI protection in Korea, along with relevant statistics, aimed at equipping foreign people with the necessary insights to navigate this evolving landscape.

Overview

According to recent data, South Korea has seen a remarkable increase in AI-related patent applications. In 2022, out of 24,341 patent applications related to Fourth Industrial Revolution (4IR) technologies, 8,960 were AI-related. This represents a substantial portion (36.8%) of the overall patent filings, underscoring the country's strong focus on AI technology. From 2013 to 2022, South Korea experienced significant growth in AI-related patents, with a notable surge in filings from 2018 to 2021. Machine learning is the dominant technique in these patents, playing a critical role in the advancement of AI technology, with applications spanning various sectors, including autonomous driving and intelligent robots.

The protection of AI in South Korea encompasses various aspects of intellectual property law, including patents, copyrights, trade secrets, and data protection. The Korean Intellectual Property Office (KIPO) plays a crucial role in administering these protections, ensuring that AI innovations are adequately safeguarded.

Patents and AI

Patentability of AI Inventions in Korea

The Korean Intellectual Property Office (KIPO) has established specific guidelines for examining patent applications related to artificial intelligence (AI) technologies. These guidelines, released in January 2021, address the unique challenges posed by AI inventions and provide a framework for assessing their patentability.

Key Patentability Criteria

AI inventions in Korea must meet the standard patentability requirements of novelty, inventive step, and industrial applicability. However, KIPO's guidelines emphasize additional considerations:

Enablement Requirement

The patent application must provide a clear and complete description that allows a person of ordinary skill in the art (POSITA) to practice the invention. This includes:

  • Detailed Explanation: Provide a thorough explanation of the technical problem being solved and the corresponding solution.
  • Technical Configurations: Include specific details about the technical configurations, such as training data, preprocessing methods, and model architecture.
  • Implementation: Describe how the AI model is trained and implemented, ensuring that the process is clear and replicable.

Inventive Step

KIPO warns that simply mentioning the use of AI technology is unlikely to meet the inventiveness requirement. The application should:

  • Distinguishing Features: Specify distinguishing technical configurations that set the invention apart from prior art.
  • Technical Effects: Elaborate on technical effects that go beyond conventional AI technologies, providing concrete evidence of innovation.
  • Avoid Generic Claims: Avoid generic claims about improved speed or accuracy without providing detailed substantiation.

Categories of AI Inventions

KIPO's guidelines recognize different categories of AI inventions, each with specific considerations:

  1. AI Training Data Inventions:
    • Focus on data preprocessing and feature extraction.
    • Describe the correlation between raw data and training data.
    • Explain the rationale behind specific preprocessing techniques.
  2. AI Modeling Inventions:
    • Detail the technical configuration of the model architecture.
    • Specify training processes and methods.
    • Highlight novel aspects of the model design or training approach.
  3. AI Application Inventions:
    • Provide information on the relationship between input and output data.
    • Describe how the trained model solves a specific technical problem.
    • Explain the practical application and benefits of the AI system.

Drafting Strategies

To improve the chances of patent grant, applicants should:

  • Articulate the Technical Problem: Clearly define the technical problem being addressed by the AI invention.
  • Provide Specific Examples: Include detailed examples of data preprocessing, model architecture, and training methods.
  • Differentiate from Conventional Approaches: Explain how the AI invention differs from existing technologies and methods.
  • Substantiate Improvements: Quantify and substantiate any claimed improvements in performance or efficiency.
  • Describe Practical Applications: Detail the practical applications and real-world impact of the AI technology.

By following these guidelines, applicants can better position their AI inventions for successful examination by KIPO. It's important to note that the field of AI patent law is evolving, and practitioners should stay informed about the latest developments and examination practices at KIPO.

Copyright and AI

Protection of AI-Generated Works

Currently, Korean copyright law primarily protects works created by human authors. However, discussions and potential legislative changes aim to recognize and protect AI-generated works, particularly where significant human involvement is present.

Trade Secrets and AI

AI technologies, including algorithms, data sets, and models, can be protected as trade secrets if they remain confidential, have commercial value, and are subject to reasonable measures to maintain secrecy. South Korea offers robust legal mechanisms for enforcing trade secrets, including civil and criminal remedies.

Data Protection and AI

The use of personal data in AI systems is regulated under the Personal Information Protection Act (PIPA). Key considerations include obtaining informed consent, data minimization, and anonymization where possible. AI developers must implement robust data security measures to protect personal and sensitive data used in AI systems.

Innovation and Future Outlook

Government Initiatives

The South Korean government supports AI innovation through various initiatives, including R&D funding, regulatory sandboxes, and promoting collaboration between academia, industry, and government.

Future Trends

South Korea is integrating advanced technologies into its IP system, focusing on legislative updates, international cooperation, and developing ethical guidelines for AI development and use.

Conclusion

The protection of AI in South Korea is comprehensive, covering various aspects of intellectual property law. Understanding these protections is crucial for foreign patent attorneys to effectively navigate the Korean IP landscape. With ongoing government initiatives and potential legislative changes, South Korea remains a dynamic and attractive jurisdiction for safeguarding AI innovations.