Somewhere inside the Korean Intellectual Property Office in Daejeon, an examiner is reading your claim language for the very first time, and they will read it exactly once before deciding whether it clears the bar for clarity. There is no second impression. If the wording is vague, over-broad, or internally inconsistent because something was lost in the handoff from English to Korean, that examiner isn’t going to guess at what you meant — they’re going to issue an office action and start the clock ticking on months of avoidable back-and-forth. 📄
This is the reality of filing in one of the world’s busiest patent jurisdictions. KIPO has ranked among the top five patent offices globally for years, receiving 237,633 patent applications in 2022, climbing to 243,310 in 2023Intellectual Property Office (KIPO) (237,633) and the EPO (193,610)the Japan Patent Office (JPO) (300,133), the Korean Intellectual Property Office (KIPO) (243,310), and reaching 246,245 in 2024followed by the Japan Patent Office (JPO) (306,855), the Korean Intellectual Property Office (KIPO) (246,245) and the European Patent Office (EPO) (199,402). That volume places enormous pressure on examination timelines, and it means every examiner is triaging a heavy caseload where clarity is the difference between a smooth path to grant and a prolonged fight. World Intellectual Property Indicators 2023 +2
The Math Behind the Rejection Rate
Numbers tell the story better than adjectives do. In 2024 alone, KIPO logged 199,407 requests for substantive examination, issued 186,382 preliminary rejections, and finalized 39,761 outright refusals, while ultimately allowing 122,382 applicationsIn total, 199,407 requests for substantive patent examination were made in 2024, and the KIPO issued 186,382 preliminary rejections and 39,761 final rejections, while allowing 122,382 applications during the same period.. That puts the overall allowance rate at 74.6 percent, a figure that has climbed steadily from 60 percent back in 2016 but dipped slightly to 72.9 percent in 2023 before recoveringThe allowance rate of patent cases was 74.6%. The allowance rate had been steadily increasing from 60.0% in 2016 and peaked at 74.3% in 2022, before a slight decrease to 72.9% in 2023.. ApaaonlineApaaonline
| Metric (2024) | Figure |
|---|---|
| Substantive examination requests | 199,407 |
| Preliminary rejections issued | 186,382 |
| Final rejections | 39,761 |
| Applications allowed | 122,382 |
| Overall allowance rate | 74.6% |
| Appeal success rate | 28.1% |
Look closely at that first row against the third: nearly every application that eventually gets granted first passes through a preliminary rejection stage. Some of those rejections are substantive — novelty, inventive step, prior art. But a meaningful share trace back to something far more preventable: descriptions and claims that don’t hold together cleanly once translated, forcing the examiner to flag ambiguity before they can even evaluate patentability on the merits.
Foreign applicants are not a rounding error in this system either. Roughly one in five patent filings at KIPO comes from outside Korea, with U.S., Japanese, German, Swiss, French, and British applicants making up the largest share of that groupFor patents, about 79.5% of the applications were filed by Korean applicants, while 20.5% were filed by foreign entities. The top seven foreign nationalities among the foreign applications, together their respective number of applications filed, were: the United States (15,576), Japan (13,861), China (5,652), Germany (3,650), Switzerland (1,639), France (1,475) and the United Kingdom (1,318).. Every one of those applications had to cross a language boundary before an examiner ever laid eyes on it, and Korean is consistently ranked among the top languages used in official PCT publications, underscoring just how much cross-border patent traffic runs through Korean-language filings 🌏. Apaaonline
What “Ambiguity” Actually Means to an Examiner
It rarely looks like a typo. More often, it’s a single term that could plausibly mean two different things once it crosses into another language, and the examiner has no way to know which one the original inventor intended. Consider a device described in the source application as a portable medical monitor. In the source language, that phrase might clearly imply something a patient wears or carries. Translated into another jurisdiction’s language without technical precision, the same term can broaden enough to cover a stationary hospital-grade monitor instead — an entirely different scope of protectionSuppose your English application describes a “portable medical monitor.” In one country, the translated term may clearly mean a device worn or carried by a patient. In another country, the term may be interpreted broadly enough to include a large hospital monitor. That difference could affect the prior art the examiner considers and the products covered by the claim.. That single word choice can change which prior art the examiner pulls, and it can change what a competitor is legally allowed to build around your patent later. Nolan Heimann
This is precisely why KIPO’s own examination guidelines explicitly flag the risk of inconsistency between the Korean-language and English-language versions of examination standards, treating that gap as something examiners are trained to watch for rather than dismissor ambiguity between the Korean edition and the English edition. If the office’s own internal guidance material draws a hard line around cross-language consistency, it’s a reasonable signal for how strictly the same standard gets applied to the claims and specifications sitting on an examiner’s desk every day. Kipo
Machine translation tools compound the risk rather than solving it. They’re genuinely useful for getting the gist of an office action or a foreign filing quickly, but patent language is deliberately dense with legally loaded terms, and that density is not always accidental — it’s often the product of years of claim drafting strategyMachine translation and artificial intelligence can be useful for understanding a foreign document. They can help you get the general idea of an office action, a patent, or a letter. But using machine translation as the final version of an official patent filing can be risky. Patent language is full of ambiguity, and the ambiguity is not always accidental.. An automated tool has no way to know that “housing” in claim 3 needs to match, word for word in legal effect, the same structural element described three paragraphs earlier in the specification. Nolan Heimann
The Amendment Trap
The risk doesn’t disappear once the application is filed — it resurfaces the moment a rejection triggers an amendment. If an examiner cites prior art disclosing a sensor, and the applicant wants to narrow the claim by specifying that the sensor sits inside a particular housing, that new language has to land in Korean with the exact same technical meaning as the original English disclosure. Get the wording even slightly off, and the amendment can look like it introduces new matter that wasn’t supported in the original filing — a problem that can undermine the entire prosecution strategy built around that claimThat amendment must be expressed accurately in the foreign language and supported by the original application. If the translation makes the housing sound different from the housing described in the original English application, you could face a support or added-matter problem.. Nolan Heimann
This is where timelines get expensive. KIPO’s average first office action pendency currently runs around 16.1 months for patents and utility modelsThe annual average first office action pendency period was 16.1 months for patents and utility models. — over a year before an applicant even learns whether the examiner had concerns. Add a translation-driven ambiguity issue into that mix, and the response cycle stretches further, pushing grant dates back and, in fast-moving technology sectors, potentially handing competitors a window to design around an unresolved claim. Hbccpit
For applicants who do end up contesting a rejection, KIPO’s Trial Board reviewed 1,558 appeal cases in 2024, with a 28.1 percent success rate — a modest improvement from 27.1 percent the year before1,380 trial petitions were appeals against the examiner’s final rejections and a total of 1,558 appeal cases were reviewed by the Trial Board. Among the cases reviewed, 438 patent applications were either allowed at the Board’s discretion or remanded to the Examination Bureau for re-examination. The success rate of appeal in 2024 (28.1%) increased slightly from 2023 (27.1%).. That’s not a coin flip you want to be relying on when a cleaner translation at the filing stage could have avoided the appeal altogether. Apaaonline
Why the Volume Makes Precision Non-Negotiable
Between 2022 and 2024, KIPO’s patent filing volume grew even as global filing patterns shifted, and Korea remained one of the consistent drivers of that growth alongside China and IndiaA substantial rise in filings by applicants residing in China, the Republic of Korea (+15,628), the United States of America (US), Japan and India were the main driver of growth in 2023.. PCT filings routed through KIPO as the receiving office also ticked upward, from 21,916 in 2022 to 22,166 in 2023The number of PCT applications filed to KIPO increased by 1.1 percent from 21,916 in 2022 to 22,166 in 2023., adding yet another layer of translated documentation that examiners have to reconcile against original filings. wipoHbccpit
When an office is processing this many applications every single year, examiners simply don’t have the bandwidth to interpret a poorly worded claim charitably. They read the document as written, apply the statutory clarity requirement, and move on. The applicant then absorbs the cost of that ambiguity in the form of extended prosecution, additional attorney fees, and — in the worst cases — a narrower claim scope than the invention actually deserved.
The Practical Takeaway for Applicants Filing Into Korea
None of this means Korean patent prosecution is unusually harsh — it means it’s unusually literal. Claims are read exactly as they’re written, and the gap between “close enough” and “technically precise” is where most avoidable rejections live. Applicants who treat the Korean translation of their specification as a formality rather than a technical-legal document are the ones most likely to end up on the wrong side of a preliminary rejection notice, an amendment dispute, or an appeal with roughly a one-in-four chance of success.
Filing into a jurisdiction that processes this many applications with this little tolerance for loose wording is not the place to treat claim translation as an afterthought. The examiner reads it once, decides once, and the cost of getting that first read wrong shows up months later in the form of an office action nobody budgeted for. 🇰🇷
Further reading on Korea’s patent examination environment and global filing trends is available directly from the source organizations:
– WIPO World Intellectual Property Indicators (wipo.int)
– KIPO official statistics and examination guidelines (kipo.go.kr)
– APAA Online industry updates (apaaonline.org)