SUWON — The Gyeonggi Venture Association (GIVA) and Hypernology jointly held the 6th G-Venture Management and Knowledge Seminar on the 22nd at the Gyeonggi Business & Science Accelerator (GBSA) in Suwon, under the theme of protecting venture companies' core assets and managing business risks in the age of artificial intelligence.
The seminar, attended by venture entrepreneurs in the province including Hypernology CEO Lee Nam-hee, was organized as the spread of AI technology accelerates competition in patent filings and raises the importance of managing labor issues and trade secrets within companies. It aimed to strengthen venture firms' risk response capabilities by providing training on key management issues needed at each stage of corporate growth, including patents, labor affairs and the defense of core technologies. The event was held free of charge for full member companies of the association that had registered in advance.
In the first session, patent attorney Hwang In-jin gave a lecture on AI patent strategies for venture companies. Hwang said AI-related patent filings increased sharply around 2023 and 2024. He added that statistics based on the Korean Intellectual Property Office classification codes alone make it difficult to determine the overall trend, and that it typically takes about 18 months from filing to publication, which must also be taken into account.
He stressed that because Korea's patent system follows the first-to-file principle, simply having invented something first does not guarantee rights. If a competitor files for the same technology first, latecomers can be put at a disadvantage, he said. Accordingly, he suggested that early-stage companies secure breadth in their portfolios through preemptive filings, while filing separately for core technologies to raise the quality of their rights.
He also explained that for inventions involving AI, records proving human contribution are important. Researchers should document what conditions they designed and which AI outputs they selected, modified or verified, keeping research notes and records of prompts, he said. In particular, he cautioned that in the fields of chemical compounds and pharmaceuticals, AI-generated data must not be submitted as if it were actual experimental results.
In the second session, attorney Kim Sung-hoon addressed labor risks and strategies for defending core technologies and trade secrets. Kim stressed the importance of procedures and records throughout hiring, evaluation, disciplinary action, dismissal and departure. Employment contracts must clearly state wages and working conditions, he said, adding that even freelance contracts can be deemed employment relationships if actual direction and supervision are involved.
Managing underperforming employees and dismissal were also presented as major issues. Kim said subjective dissatisfaction alone is unlikely to justify a dismissal. Evaluation criteria, opportunities for improvement, procedures allowing the employee to respond, and specific grounds must be documented in order to respond to disputes, he advised.
On trade secret protection, he explained that non-disclosure, economic value and secrecy management are all required. Material does not become a trade secret simply because it is internal company data, he said, adding that access restrictions, confidentiality markings, approval for removal and log management must actually be in operation. He also said companies should establish internal standards to prevent the indiscriminate input of personal data, non-public source code, manufacturing formulas and API keys into external AI services when using generative AI.
