AI patent disputes up ninefold, but patent exam has no AI subject — and no passers

As AI-related patents and disputes multiply, the patent attorney licensing exam contains no dedicated subject to assess AI expertise, it has emerged. The Intellectual Property Office considers data structures — a course already included in AI and software engineering curricula — sufficient to evaluate such competency, but the number of exam candidates choosing that subject has fallen to roughly one-tenth of its 2017 level, even as AI-related patent tribunal filings surge.

According to the office of People Power Party lawmaker Jo Ji-yeon, a member of the National Assembly's Industry, Trade, Resources, SMEs and Startups Committee, the current patent attorney exam offers no elective subjects covering AI, machine learning or software engineering.

The Intellectual Property Office has explained that data structures is already a required course in AI and software engineering degree programs, and that the subject can therefore serve as a proxy for evaluating expertise in those fields — making a separate AI elective unnecessary.

The problem is that demand for AI patent work is rising while the number of exam candidates selecting data structures is falling. According to data submitted by the Intellectual Property Office, only five candidates chose data structures as their second-round elective last year — and not one passed.

In 2017, 48 candidates sat the data structures elective and six passed. By 2024, those figures had dropped to 10 candidates and one passer; last year the number of candidates fell again by half, to five. Compared with 2017, the data structures candidate pool has shrunk by roughly 90 percent in eight years.

Meanwhile, patent disputes and tribunal demand surrounding AI technology are growing rapidly. The Intellectual Property Office has operated a dedicated AI tribunal division since May 2024. AI-related tribunal filings stood at just three cases in 2024, then jumped to 27 last year — a ninefold increase in a single year. Fifteen more cases were filed through July this year.

The office has already stood up a specialized internal unit to handle the rising caseload as AI technology spreads, yet the patent attorney exam still has no subject dedicated to assessing AI expertise, and candidates for data structures — the subject the office points to as the closest equivalent — are disappearing.

The Intellectual Property Office said it is addressing the expertise gap by assigning specialist staff and providing dedicated training. "We have placed AI doctoral holders and senior examiners and adjudicators in the division, and we are supporting continuous updates on fast-moving technologies through seminars with outside experts and external professional training," the office said.

For practicing patent attorneys, AI-related competency is being supplemented through mandatory continuing education separate from the licensing exam. Training offered through the Korean Patent Attorneys Association's mandatory refresher program covers AI examination practice, cutting-edge AI-related technologies, AI-assisted patent work, and AI use and information security.

Efforts are also under way to strengthen education in response to the spread of generative AI. "Since last year, we have been in ongoing discussions to reinforce training so that patent attorneys are aware of information security and related issues when using generative AI in their work," the office said.