|
|
|
|
 |
|
 |
|
HOME > Resources >
Newsletters |
|
|
| |
|
|
Title |
|
MOIP opens 3 types of AI learning data that are difficult for companies to directly build on their own.(News Letter No. 588) |
|
|
Author |
|
KH |
|
Post Date |
|
2026-11-16 |
|
|
Read |
|
43 |
|
Attach File |
|
- |
|
1. MOIP opens 3 types of AI learning data that are difficult for companies to directly build on their own.
MOIP announced to open freely 3 types of intellectual property (IP) data for AI learning, about 9,610,000 cases, through the intellectual property information service (KIPRISPlus Service).
Since IP data, such as patents and trademarks, etc., include a lot of technical terms and have a complex document structure, it requires considerable time and cost to process them into forms suitable for AI learning. Notably, it is a significant burden for small and medium companies or startups, etc. to directly build large-scale learning data on their own.
The data newly opened at this time are 3 types: ►patent summary learning data, ►experimental learning data in the chemical and metal fields, and ►learning data for product classification models, which were built through the ‘AI-based patent administrative innovation project’ of MOIP. When these 3 types are added, the AI learning data to be opened totals 22 types and increases to about 57,000,000 cases. The patent summary learning data (about 16,000 cases) are a dataset where experts have summarized the core content of patent specifications, such as the ‘technical problem’, ‘representative claim’, and ‘effects of the invention’, etc., and paired the original texts and their respective summaries. These data can be used to develop the AI patent summary models to evaluate the performance and quality of summaries. The experimental learning data in the chemical and metal fields (about 3,000,000 cases) are converted and structured data of experimental result tables included in patent documents in the form of images or text, so as to be analyzed by a computer by using AI and optical character recognition (OCR) technology. The experimental information that was previously difficult to be utilized because it was in the form of image can be now used for research and development, industrial analysis and patent searches, etc.
The learning data for product classification models (about 6,600, 000 cases) are the data which connect the names of products and services described in trademark applications with the similar group classification information used during the examination process, and these data can be utilized to develop AI models for automated product classification, similar product searches, and market and application trend analysis, etc. Notably, this opening is significant in that data built up and accumulated during administrative processes are converted into learning data for the private sector to use for AI development, to be released again. Since companies can use these data for AI research and development and service development, without directly collecting or processing data with a lot of time and money, it is expected to reduce private sectors’ AI development costs and help create new IP services.
2. A ‘support plan for innovation in IP-based R&D-IP-Commercialization’ was announced (September 30, 2026).
- The internalization of IP strategies in national R&D and the introduction of an AIxIP-based high-efficiency R&D models are promoted.
- International monetization is improved by supporting the securing of overseas core patents for public technologies and strengthening the commercialization capabilities of technology licensing offices (TLOs).
The support plan or innovation in IP-based R&D-IP-Commercialization which was prepared MOIP and the Ministry of Science and ICT (MSIT) was resolved at the 13th Ministerial Meeting for Science and Technology on September 30, 2026.
The number of international patent applications in the national strategic technology field increased by about 3 times, from 61,560 in 2014 to 178,791 in 2023, and the global competition for technological hegemony is intensifying into a race to secure core technologies through patents.
To actively respond to the expanding international technology hegemony competition into a patent competition, the Korean government prepared a plan to support full-cycle national R&D innovation based on IP strategies, by actively utilizing IP strategies from the early stages of national R&D to secure research outcomes as core overseas patents and connect them to commercialization.
The government carries forward 3 major strategies of: ① internalizing IP strategies in national R&D, ② securing high value-added overseas and standard patents and advancing their utilization, and ③ creating an innovative ecosystem for public IP commercialization.
Global patent trends and promising technologies in the national strategic technology field are to be analyzed to be reflected in future R&D directions and implementation roadmaps. Starting with the establishment of the ‘national patent strategy roadmap’ in the physical AI field in the second half of this year, it will be expanded to other strategic technology fields.
IP-R&D will be first applied to core R&D tasks with a strong need to preempt international IP, and then gradually expanded. Patent information will be processed into AI learning data, and an ‘AI-IP solution’ will be developed to provide optimal research design and virtual experiments and strategies for securing core IP which use the learning data, to be applied to research sites to improve the efficiency of the entire R&D processes.
Design of rights which are usable in the global market, entry into core countries, and reinforcement of rights are to be supported so that, in the strategic technology field, excellent technologies requiring preemptive overseas patenting are selected to be secured as ‘international K-IP core patents.
In the future promising fields, such as AI and robots, etc., strategies to secure standard patents from the R&D stage are implemented. Patent infringement risk assessment and response strategies thereto, and core patent acquisition are supported by comparing and analyzing overseas AI patent pools and South Korean AI companies’ patents. In the humanoid robot field, a ‘K-patent pool’ to jointly utilize Korean patents is also established gradually.
Patents held by universities and public research institutes are also evaluated to select those with high practical value. Outstanding patents are connected to demonstration, advancement, and technology transactions, and patents for innovative technologies with increasing future value are strategically supported to be retained.
Common standards for the tasks of technology licensing offices (TLOs), from invention interviews to patent applications and portfolio building and commercialization, are established, and customized support is also provided in accordance with TLOs’ capabilities and developmental stages.
‘IP Specialists’, which are dedicated IP specialists for TLOs and research project groups, are systematically trained. Through an ‘IP-TLO alliance’, excellent cases in overseas patent applications, startups, and international licensing, etc. are shared, and cooperation are strengthened. Customized educations for universities, companies and research institutes are expanded to help researchers and practitioners understand IP strategies and utilize them in R&D, patent management and commercialization.
Through this plan, the South Korean government plans to improve the efficiency of R&D, secure core IPs in the strategic technology field, and build a virtuous cycle to connect outstanding research achievements to overseas original and standard patents with international commercialization, thereby supporting innovation and advancement throughout the R&D cycle.
3. South Korea ranked 1st in the number of global patent applications related to grid-scale energy storage system (ESS) over the past 20 years.
- 4 Korean companies, including LG Energy Solution ranked 1st, Samsung SDI ranked 3rd, etc., were included in the top 10.
- South Korea took the lead in the number of the patent applications related to ESS enhancing grid flexibility.
In the number of patent applications in the grid-scale energy storage system (ESS), South Korea (by nationality) with 6,030 applications ranked 1st over the last 20 years. 4 Korean companies, including LG Energy Solution (4,605) raking 1st among companies with the most patent applications, were included in the top 10, taking the lead in filing the patent applications in this field.
MOIP analyzed the patent applications related to the grid-scale ESS, which were filed with IP5 (in South Korea, the US, China, Europe and Japan) over the last 20 years (2004~2023). As a result, the number of the relevant patent applications in this field increased from 32 in 2004 to 4,010 in 2023, with a growth of about 125 times.
While the importance of stable power supply has been increasing with the spread of AI and the growth of the advanced industries, such as semiconductors, etc., the need to respond to fluctuations in generation volume by region and time of day has been also increasing with the expansion of renewable energy, such as solar and wind. Accordingly, the grid-scale ESS is also becoming important to flexibly operate the grid by not only expanding power facilities, such as power plants and transmission lines, etc., but also easing the supply-demand gaps across power surplus or power shortage times and regions.
By applicant nationality, South Korea accounted for the largest number at 6,030 cases (41.3%), followed by China with 5,176 (35.4%), the US with 1,283 (8.8%), Japan with 1,220 (8.4%) and Germany with 250 (1.7%). In the average annual increase rate over the last 3 years (2021~2023), South Korea was found to maintain its lead at 68.6%, ahead of the US (53.1%), China (48.2%), Germany (33.3%) and Japan (12.1%). <4 Korean companies including LG Energy Solution ranking 1st and Samsung SDI ranking 3rd, etc. were included in the top 10.>
By applicant (company), LG Energy Solution ranked 1st (with 4,605 cases), CATL 2nd (2,653), and Samsung SDI 3rd (242). In addition, LS Electric (114) and Samsung Electronics (78) ranked 7th and 8th, respectively, showing that 4 Korean companies among the top 10 leaded the technology development in this field.
By specific technology, storage safety and management technologies of monitoring the storage condition and the thermal and fire risks accounted for the largest number at 7,705 cases (52.8%), followed by the long-duration energy storage technology of storing and supplying power for long periods with 5,338 (36.5%) and power conversion and interconnection technologies of converting stored power and connecting it to the grid with 1,562 (10.7%). This shows that the development of the grid-scale ESS technology has been carried out extensively, from ensuring safety to long-duration storage and grid connection.
4. Korea Intellectual Property Trial and Appeal Board (KIPTAB) launched an ‘intellectual property (IP) trial system research council’.
KIPTAB of MOIP and Korea Patent Attorneys Association (KPAA) jointly held the first ‘IP trial system research council’ and discussed major improvement tasks of the patent and trademark trial systems and measures to improve trial practices.
The IP trial system research council is a regular research group jointly arranged by the two KIPTAB and KPAA, to reflect the voices of those directly using the trial systems into trial policies. In this council, about 15 people attended, including the heads of the two institutions, officials from KIPTAB and patent attorneys affiliated with KPAA, etc.
In the first council, KPAA proposed improvement tasks raised in the field of: ►introducing a system to restrict the exercise of duplicate rights when multiple patents are registered for a single invention, and ►establishing standard decision deadlines for the case where a response was not submitted in a trademark cancellation trial for non-use trademark. KIPTAB gathered opinions from experts and the field regarding: ►2026 trial manual amendment focusing on trial handling standards for AI-generated false materials, and ► improvement plan for attorney fees in trial costs.
The council plans to discuss the major issues, such as enhancing the reliability and predictability of trial procedures, improving the system of requesting patent cancellation, and responding to changes in the trial environment due to the expansion of AI utilization, etc. KIPTAB and KPAA will also regularly operate the council, share the opinions from the field, and continue to identify improvement tasks to enhance the expertise and speed of trials and the user convenience.
The president of KPAA said, “the research council has great significance in that patent attorneys directly share their trial practice experiences and improvement suggestions with KIPTAB… We will positively cooperate to establish more rational and user-friendly trial systems based on the field experiences.” The president of KIPTAB said, “it is important to continue to listen the voices from the field, to development the trial systems so as to be effective means of remedying rights that citizens and companies can trust…We will closely cooperate with KPAA so that the opinions discussed in the council can lead to actual system improvements.”
5. 3 former employees were caught for leaking core semiconductor inspection technology overseas.
- They committed at double salary offers… one former employee was arrested and the other two were indicted without detention.
According to MOIP’s special judicial police division for technology leakage (hereinafter, referred to as the ‘technology police’) and Daejeon District Public Prosecution Office’s major crimes indictment division 2, former employee “A” (45 years old), who used the victim company’s core semiconductor inspection technology for overseas production, was arrested and indicted on charges of violating the Unfair Competition Prevention Act and the Industrial Technology Protection Act, and former employees “B” (40) and “C” (40) were indicted without detention.
<3 former employees were arrested for using, overseas, the core semiconductor inspection technology successfully localized by South Korea. >
According to the technology police, “A” and others have been accused of conspiring with an overseas distribution company responsible for local distribution of the Korean victim company from November 2023 to February 2026 to establish an overseas manufacturing company, producing products using the design drawings and data of the ‘core semiconductor inspection technology’ of the Korean victim company and supplying them overseas.
As a result of investigation, it was revealed that “A”, “B” and “C” received double salary offers from “D” (58), who had fled overseas, and changed jobs to the overseas manufacturing company (November 2023).
It was found that they supplied the prototypes manufactured by using the victim company’s core semiconductor inspection technology (probe card technology) overseas (December 2025), and the victim company estimated that if it lose overseas markets due to the suspects’ crimes, an annual revenue loss of 40 billion Korea Won would occur. The probe card technology relates to the semiconductor inspection equipment to inspect whether a semiconductor chip is defective in the wafer state, and this is a field that has been dominated by the US and Japanese companies. The Korean victim company succeeded in localization and even developed the equipment for high-bandwidth memory (HBM) inspection, supplying products to major domestic and overseas semiconductor big companies.
The technology police started an investigation in July 2025 after getting intelligence from the National Intelligence Service (NIS) but faced difficulties as the suspects including “A” were staying overseas.
The technology police in cooperation with NIS grasped the situation indicating that an overseas manufacturing company established by the suspects used the Korean domestic supply chain for procurement and design and then, searched and seized the domestic design company (July 2026), securing the victim company’s blueprints and technical data used by the suspects.
Since the sources, such as the victim company’s name, etc., were deleted in the materials secured by the search and seizure, or the materials were reedited to look like documents belonging to the overseas manufacturing company to which the suspects transferred, it was difficult to determine whether they were the victim company’s technical data solely based on their appearance.
The technology police intensively deployed professional investigators with R&D experience in the semiconductor industry, along with additional investigative personnel added (June 2026). By precisely analyzing the searched and seized materials, it was also revealed that the materials with deleted sources were written and altered based on the victim company’s technology.
The technology police referred “A” to the prosecution with a recommendation for indictment under detention, and “B” and “C” with a recommendation for indictment without detention. As for the prime suspect, “D” (former executive of the victim company), who is currently staying overseas to evade the investigation, an Interpol Red Notice has been requested and the investigation will proceed immediately once he is in custody.
|
|
|
|
|
Prev |
|
Korean IP News No. 157 in Chines
|
|
|
Next |
|
Korean Industrial Insight No.32 of AI, Semiconductors, Batteries, Electric vehicles, Robot, and Bio in chinese
|
|
| |
|
|
|
|
|
|
| |
|
|
|
|
|
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|