Korean SPF and US sunscreen labels: what should AI answers say?

Nothing stops a label from travelling. A brand publishes its Korean packaging copy in English, a retailer copies the page, an aggregator copies the retailer, and an assistant quotes whichever of the three it retrieved. By the time a US buyer reads the sentence, it has been repeated by three sources that all got it from the brand — and none of them checked whether the US product’s file supports it.

That is why this is a correction problem before it is a compliance problem. An assistant will not invent an SPF; it will faithfully repeat the one you published. The work is knowing which of your own claims were never portable, so you can stop them travelling and fix the ones already out there.

Sources checked on 21 September 2026. This is a publishing and evidence guide for brand teams, not a product approval or a legal opinion — the brand’s regulatory owner decides what the US product may claim.

Which Korean label claims can cross into a US answer?

None of them cross unchanged. The table below maps each Korean claim to what it does not establish about the product sold in the US, and to what a US claim would need instead. Every US rule below sits in 21 CFR 201.327, the sunscreen labeling regulation, unless another source is named — current text, or the 2024 annual edition as a PDF.

The second row is the one that catches brand teams. A US front-of-pack statement is a single composite — “Broad Spectrum SPF 50” — and its two halves come from two different tests: the number from paragraph (i), the words from paragraph (j). There is no four-plus scale anywhere on a US label for PA++++ to land on. Dropping it understates a product that may well pass; translating it as “Broad Spectrum” asserts a test result nobody has shown you. Find out whether the US product passed the paragraph (j) test, then say that and nothing more.

The third row is the one Korean teams read past. 기능성화장품 is a cosmetics category reviewed by MFDS, and SPF and PA data are among the materials submitted for it. The US does not have that category for sunscreen at all: it regulates these products as drugs. A completed Korean review is real evidence about a real product — it is simply evidence in the wrong jurisdiction for a US sentence.

The sections below take the rows that need explaining — SPF, PA++++, and the two approval claims — and then turn to correcting an answer that already carries one of them.

Does Korean SPF mean something different from US SPF?

SPF describes protection against sunburn; a matching number does not establish that two products have identical formulas or satisfy the same market’s rules. Check the exact product and its supporting test record before comparing claims. The FDA’s sunscreen explanation describes SPF as primarily an indication of UVB protection, rather than a multiplier for hours in the sun.

Korea’s MFDS functional-cosmetics guidance includes products protecting skin from ultraviolet rays and identifies SPF and UVA-protection data among the supporting materials. The United States regulates sunscreens as drugs. These are separate regulatory frameworks, even when a brand uses similar packaging in both markets.

For a useful product comparison, request the Korean and US product names, formula or SKU identifiers, ingredient lists, label images, test methods and test dates. Put the market beside every value. If the brand cannot confirm that a test report belongs to the formula being described, leave that comparison unresolved instead of borrowing a number from another product.

Can we translate PA++++ into US Broad Spectrum?

Do not replace PA++++ with Broad Spectrum merely by translating a Korean label. Each claim stays tied to the market it was tested in: US Broad Spectrum labeling depends on meeting the applicable US test. 21 CFR 201.327(a), (i) and (j) distinguishes the SPF procedure from the broad-spectrum procedure (current text, PDF).

Under that section, the broad-spectrum test uses a critical wavelength of at least 370 nanometers. Water resistance is another separate claim: the prescribed statements specify 40 or 80 minutes according to testing. The section also identifies “waterproof,” “sweatproof” and “sunblock” as misleading claims. A translation should preserve those distinctions, not add an assurance the test record does not support.

Use a small record for each market: the exact label claim, applicable test, tested formula, document date and person who confirmed the wording. Publish a readable summary where disclosure is permitted. A private report can support internal review without being uploaded publicly. Mark an unavailable report as unavailable; do not turn that gap into a claim that testing never happened.

Does Korean approval mean a sunscreen can be marketed in the US?

No. Korean regulatory status does not establish compliance with US sunscreen requirements. Record the market and its documentation separately; the brand’s regulatory specialist determines whether the specific US product meets the applicable requirements. MFDS’s process concerns Korean functional cosmetics, while FDA’s explanation describes how qualifying OTC monograph drugs can enter the US market without an approved drug application.

That distinction matters when writing “FDA approved.” Meeting monograph conditions is different from holding an approved drug application for the product. Ask the compliance owner which route applies and which wording the documentation supports.

Check the current rule instead of copying an old filter comparison. FDA announced bemotrizinol’s addition in June 2026; final order OTC000039 specifies conditions, an effective-date provision and exclusivity. It does not make every imported formula containing that ingredient compliant. FDA’s September 2026 regulatory update also separates finalized provisions from proposals still under review. Do not describe a proposal as an effective requirement.

For publishing, keep three fields distinct: what the package says, what the test demonstrates, and what the regulatory review permits. Give the reviewer the actual label and formula, not just a brand name or retailer screenshot.

How should a brand correct an AI answer about its sunscreen?

Start with the exact answer, the pages it cited and the product record for the market the question named. A correction should identify the wrong statement, provide the approved replacement and link the evidence for that replacement. Who is able to act on each one — the brand, a retailer, or the engine — is a separate question.

1. Save the prompt, engine, date, full answer and every cited URL. Note which market the question specified.
2. Identify the exact disputed claim: formula, SPF, broad-spectrum status, water resistance or approval wording.
3. Have the compliance owner confirm a replacement using the relevant label and test records.
4. Update the official product page and request matching corrections from retailers carrying the error. Keep the previous and replacement text in the record.
5. Submit provider feedback where appropriate, using the reporting checklist, then repeat the original question and save the new answer separately.

Then wait, and ask again. A published correction and a changed AI answer are two separate events, and the gap between them is the part most teams underestimate: the page can be right for weeks while the answer is still wrong, because the engine has not fetched it again, or is still reading a retailer that never updated. Record both dates. If the answer still repeats the error, the source it is actually reading has not been found yet.