Original Research · August 2026

How Often Does AI Get Beauty Certifications Wrong?

Five AI assistants were asked which beauty brands hold specific environmental credentials, in three markets and three languages. Every name was checked against the certifying body's own published register. Forty-six claims were verifiably false, and 39 of those errors were built on something true, reported at the wrong time, about the wrong entity, to the wrong degree, or from the wrong scheme.

English (US)
Français (FR)
Español (MX)

Headline findings

46
Verifiably false certification claims
39
That were built on something true
151
Made while citing the official register
12
That gave a parent's credential to a brand
Accuracy Study

Why were certification claims checked against public registers?

An independent check of AI certification claims against public registers

Counts and ordinal persistence only. No rates are projected from this sample. At three runs per prompt, existence and recurrence are reported as observed facts, never as estimated frequencies.

Research Findings

What did we find when answers were checked against certifying-body registers?

When an engine cites the certifying body's own register, does it get the answer right?

High

Eucerin was named as an EcoBeautyScore member in 24 answer cells across all five engines and all three markets. Twenty of those answers cited ecobeautyscore.com. Schauma was named in 22 cells, 19 of them citing the same source.

The EcoBeautyScore Association publishes its members and, under a separate adjacent heading, the brands that have published scores, printing each brand's parent company in parentheses. The engines retrieved that page and reported the second list as the first.

Implication: Retrieval worked and the answer was still wrong. A citation evidences that a source was consulted, not that it was understood, so citation share is not a measure of how accurately AI represents a brand.

Are the engines inventing certifications?

High

Mostly not. Of 46 false claims, 18 report a credential after it lapsed, 12 attach a parent's credential to a consumer brand, 7 get the degree or the scheme wrong, 2 name a company outside the category, and 7 are inventions.

An early pass classified four CDP claims as fabrication. Checking each against the company's own disclosure history showed three had genuinely held the credential or held a partial version. Beiersdorf, Kao and Symrise all really were CDP Triple A, years ago.

Implication: The engines are mostly telling you something that was true, and nothing in the answer tells you when it stopped being true. Staleness, conflation and overstatement have different causes and different fixes, and none is solved by more citations.

Whose credential is it?

High

Eight false claims attach a parent group's credential to a consumer brand that does not hold it. Six are Science Based Targets claims, where Chantecaille, Eucerin, Guerlain, Yves Rocher, Aveda and the Estée Lauder brand were each credited with their parent's validated targets.

Eucerin is conflated twice, on two different credentials, by two different routes. On EcoBeautyScore the engines misread a page that separates members from brands; on SBTi they simply attached the parent's targets to the parent's brand.

Implication: There is no rule of thumb. Aveda holds B Corp certification in its own right while its parent does not, and does not hold SBTi targets while its parent does. Whether a brand can hold a credential independently is a fact about the credential, not about the corporate tree.

Are engines reporting current certification status, or an older one?

High

Every engine answered the CDP Triple A question using the 2022 A List, reached through a trade article published in January 2023. CDP's 2025 A List had been public since 8 January 2026, 68 days before the prompts ran.

Beiersdorf, Kao and Firmenich were each named as Triple A holders. None appears among the 23 Triple A companies of 2025. Beiersdorf and Kao genuinely held the rating in earlier cycles; Firmenich has not existed under that name since its 2023 merger into dsm-firmenich.

Implication: A lapsed credential can outlive its own expiry in the answer layer. Only 3 of Beiersdorf's 28 mentions cited cdp.net at all, so here the register was not consulted rather than misread.

Does the same engine give the same answer twice?

High

ChatGPT stated 'Coty Inc.: Upgraded to a Triple A rating in 2025' in one run, and in another run of the same prompt correctly described an A rating in CDP's 2025 Climate Change disclosure. Both runs cited coty.com.

Coty's own announcement of 12 December 2025 records Climate A, Water B and Forests B-. An A on one of three questionnaires is not Triple A. The source containing the correct answer was cited in the run that got it wrong.

Implication: A brand checking this once had roughly even odds of seeing a flattering error or the truth, with no way to tell which it had received.

Do the engines fail in the same way across different credentials?

High

EcoBeautyScore answers cited the register in 70% of assertions and failed by conflation and staleness. EWG answers cited it in 64% and confused two different EWG programmes. CDP answers cited it in 13% and failed by staleness.

Two structurally different roads to a wrong answer: the primary source retrieved and misread, or the primary source never retrieved and trade press used instead.

Implication: Neither pattern is visible to a brand monitoring citation share, because in both cases the engines cited something plausible.

Were the engines wrong about everything?

High

No. The two largest claims under a credential we did not fully verify were spot-checked and both were correct. Aveda was named as a certified B Corp in 18 cells and has held the certification since January 2023. Clarins was named in 16 cells and Groupe Clarins was certified in 2025, with the certificate naming the Clarins and myBlend brands.

Aveda is a wholly owned subsidiary of The Estee Lauder Companies, which is not itself a B Corp. That is structurally identical to Eucerin being credited with Beiersdorf's membership. One is true and one is false, and nothing in either answer distinguishes them.

Implication: The finding is not that AI gets certifications wrong. It is that you cannot tell from the outside which answers are the right ones. Only the register can settle it.

Is a Skin Deep rating the same as EWG Verified?

High

No, and the engines conflated them. Era Organics was named as EWG Verified three times, all three citing ewg.org. It has 27 products rated in EWG's Skin Deep database and none carrying the EWG Verified mark.

Skin Deep is a ratings database any product can enter. EWG Verified is a paid certification requiring application, review and ongoing standards. Clean-beauty publishing uses the two interchangeably and the engines reproduced that.

Implication: A brand marketing itself as 'EWG rated' and a brand holding EWG Verified are making different claims, and the answer layer does not distinguish them.

What was the one invention?

High

A brand. Two engines named 'Natural Organic Skincare' as an EWG Verified brand in four cells, all four citing ewg.org. No brand of that name exists in the EWG Verified directory or Skin Deep; the phrase is a product category.

The only claim in the study with nothing real behind it invented the subject rather than the credential, assembling a plausible brand name out of the category language in the question.

Implication: No brand is harmed, since none exists, but a reader acting on it would be searching for a product that cannot be bought.

Can one credential be substituted for another?

High

Seven times, across four different schemes. Urban Decay and Wet n Wild hold PETA certification rather than Leaping Bunny. Saie signed The Climate Pledge rather than validating targets with SBTi. Era Organics is rated in EWG's Skin Deep database rather than EWG Verified. Coty, Shiseido and Unilever hold a single or double CDP A rather than Triple A.

Each substitutes something real, adjacent and weaker for the specific credential asked about. The engines treat 'cruelty-free', 'climate committed' and 'EWG' as single categories when each contains separate schemes with different requirements.

Implication: That distinction is the entire reason a brand pays for the stronger scheme.

How out of date can an answer be?

High

Dr. Hauschka was named as a certified B Corp in eight cells across three engines and all three markets. Its US entity was certified in September 2015 and lapsed around 2018; its German parent has never been certified.

Beautycounter is the more instructive case. Counter Brands LLC filed for Chapter 7 in April 2024, and the relaunched brand still trades, still appears on Leaping Bunny's recommitment list, and has years of sustainability press behind it. It has not recertified.

Implication: A lapse leaves almost no public trace, while the original certification leaves a great deal. The answer layer weights what is abundant over what is current.

Core Evidence

Which certification claims were verifiably wrong?

Every row is a false claim traced to the certifying body's published register. Sorted by answer-cell count as observed in the study.

CertificationBrandCellsEnginesMarketsCited registerError typeSeverityGround truth
CDP Triple ABeiersdorf28534StalenessMediumTriple A in the 2022 and 2023 cycles. Lost it in the 2024 cycle on a B for Forests while holding A for climate and A for water. Not among the 23 Triple A companies published 8 Jan 2026.
CDP Triple AKao27534StalenessMediumTriple A for five consecutive years to the 2024 cycle. Not among the 23 on the 2025 A List.
EcoBeautyScoreEucerin245320ConflationMediumListed by the Association under 'Brands which have published scores', separately from its members. Beiersdorf is the member.
CDP Triple AFirmenich23533StalenessHighMerged with DSM in 2023 to form dsm-firmenich. Neither name appears among the 23 Triple A companies in 2025.
EcoBeautyScoreSchauma225319ConflationMediumListed under 'Brands which have published scores', separately from members. Henkel is the member.
EcoBeautyScoreUnilever13537StalenessMediumNamed a founding Consortium member in NATRUE's March 2022 announcement but absent from the Association's May 2025 member list and from its 70+ contributor list.
B CorpL'Occitane12425ConflationMediumNamed as a portfolio brand in the L'OCCITANE Group B Lab profile; not separately certified.
EcoBeautyScoreThe Estée Lauder Companies12437StalenessMediumParticipated as ELC Whitman Laboratories Ltd; absent from the May 2025 member list.
EcoBeautyScoreClarins9327StalenessMediumClarins SAS is in the historical contributor list; absent from the May 2025 member list.
B CorpDr. Hauschka8337StalenessHighThe US entity was certified in September 2015 with no recertification found; the lapse is estimated at 2018 on the then three-year cycle. Parent WALA Heilmittel is not certified. Roughly seven and a half years stale at the study date.
B CorpEvolve Organic Beauty8228ConflationMediumThe Good Ventures Ltd B Lab profile names Evolve Organic Beauty as its consumer brand; the brand is not separately certified.
B CorpBYBI Beauty7317StalenessMediumCertified under founder ownership and not carried forward after the sale around 2022-23. B Lab requires an acquired company to recertify within a year.
B CorpErborian7325ConflationMediumNamed as a portfolio brand in the L'OCCITANE Group B Lab profile; not separately certified.
B CorpMustela6220ConflationMediumThe B Lab profile is titled 'Laboratoires Expanscience (main brands: Mustela and Babo Botanicals)'. The register itself states the brand is covered by the parent.
EcoBeautyScoreP&G6432StalenessMediumProcter & Gamble Services Company NV is a founding Consortium member; absent from the May 2025 member list.
Leaping BunnyL'BEL6416FabricationMediumNot found in the CCIC or Cruelty Free International Leaping Bunny registers.
Science Based TargetsChantecaille6214ConflationMediumA Beiersdorf brand since the February 2022 acquisition. Beiersdorf holds the validated targets; the brand does not report to SBTi.
Science Based TargetsPer Purr6310FabricationMediumAbsent from the SBTi Companies Taking Action dashboard, which is a comprehensive public register of validated targets.
B CorpBeautycounter5234StalenessMediumCounter Brands LLC filed for Chapter 7 in April 2024. The brand relaunched under G2G Ventures PBC, which has not recertified as a B Corp.
CDP Triple ASymrise5120StalenessMediumTriple A in 2021 (one of only 14 worldwide) and 2022. Scored A, A and A- in 2023, so not Triple A, and not among the 23 in 2025.
EcoBeautyScoreBabor5321StalenessMediumIn the historical contributor list and founding Consortium; absent from the May 2025 member list.
EcoBeautyScoreNatura &Co5321StalenessMediumNatura & Co Avon Cosmetics is a founding Consortium member; absent from the May 2025 member list.
EcoBeautyScoreOriflame5421StalenessMediumOriflame Research & Development Limited is a founding Consortium member; absent from the May 2025 member list.
CDP Triple AShiseido4213Attribute errorMediumDouble A only. Climate and Forests in the 2023 cycle; Climate and Water in 2024 and 2025. Never Triple A.
CDP Triple AUnilever4221Attribute errorMediumA for climate and forests in the 2022 cycle, not all three. Not among the 23 Triple A companies in 2025.
EWG VerifiedNatural Organic Skincare4214FabricationHighNo brand of this name exists in the EWG Verified directory or Skin Deep. The phrase is a product category, and naturalorganicskincare.com is a content site rather than a brand.
Leaping BunnyJeffree Star Cosmetics4211StalenessMediumAbsent from the 2024 and 2025 recommitment lists in either section, indicating removal before the 2024 reporting period.
Science Based TargetsEstée Lauder (brand)4120ConflationMediumThe brand is distinct from The Estée Lauder Companies, which is the SBTi-registered entity.
Science Based TargetsEucerin4210ConflationMediumA Beiersdorf brand. Only corporate groups report to SBTi; Beiersdorf holds the validated targets.
Science Based TargetsGuerlain4210ConflationMediumAn LVMH Maison. LVMH holds the validated carbon trajectory; Guerlain is not a separate SBTi registrant.
Science Based TargetsThe Today Project4210FabricationMediumAbsent from the SBTi dashboard. No commitment or validated target found.
Science Based TargetsTiffany & Co.4210Framing errorLowA luxury jeweller, LVMH-owned since 2021. It does hold validated SBTi targets, so the credential is real and the category is wrong.
Science Based TargetsYves Rocher4220ConflationMediumThe flagship brand of Groupe Rocher, which holds the validated targets. The brand does not report separately.
CDP Triple ACoty3110Attribute errorMedium2025 CDP scores were Climate A, Water B, Forests B-. An A on one questionnaire is not Triple A.
EcoBeautyScoreJohnson & Johnson Consumer Inc.3222StalenessMediumBecame independent as Kenvue in July 2023. Kenvue is a current member; the named entity no longer exists.
EcoBeautyScorePZ Cussons3210StalenessMediumPZ Cussons PLC is a founding Consortium member; absent from the May 2025 member list.
EWG VerifiedEra Organics3113Attribute errorMedium27 products rated in EWG's Skin Deep database at low to moderate hazard, but none carries the EWG VERIFIED mark. A Skin Deep rating is not the certification.
Leaping BunnyCatrice3213FabricationMediumNot found in the CCIC or Cruelty Free International Leaping Bunny registers.
Leaping BunnyEssence3213FabricationMediumNot found in the CCIC or Cruelty Free International Leaping Bunny registers.
Leaping BunnyPacifica3213StalenessMediumOn the 2019 recommitment list, absent from 2024 and 2025, and the brand page now returns 404. Pacifica's own FAQ confirms it holds PETA certification only.
Leaping BunnyUrban Decay3213Attribute errorMediumHolds PETA Beauty Without Bunnies certification, not Leaping Bunny. Absent from both the CCIC and Cruelty Free International registers.
Leaping BunnyWet n Wild3113Attribute errorMediumMarkets itself as cruelty-free through PETA only. Absent from both Leaping Bunny registers.
Science Based TargetsAveda3110ConflationMediumELC holds the validated targets. Notably Aveda does hold B Corp certification in its own right, so whether a brand can hold a credential independently varies by credential rather than by corporate structure.
Science Based TargetsAxiology3110FabricationMediumAbsent from the SBTi dashboard. No emissions reduction targets published.
Science Based TargetsSaie3210Attribute errorMediumSaie signed The Climate Pledge in 2024, the Amazon-founded net-zero-by-2040 initiative, and was the first beauty brand to do so. The Climate Pledge involves no independent target validation and is not SBTi.
Science Based TargetsTHG3110Framing errorLowAn e-commerce and technology group operating beauty retail through Lookfantastic. Holds validated targets; not a beauty company.
Error taxonomy

How do false certification claims break down by error type?

Staleness

18
claims
171answer cells

A credential reported after it lapsed, or under an entity name that no longer exists. The dominant failure. The oldest is roughly seven and a half years out of date.

Conflation

12
claims
104answer cells

A parent company's credential attributed to a consumer brand that does not hold it. Spans three credentials.

Fabrication

7
claims
29answer cells

Nothing real behind the claim. One of the seven names a brand that does not exist at all.

Attribute error

7
claims
23answer cells

A partial rating reported as full, or a real certification from a different and weaker scheme substituted for the one asked about.

Framing error

2
claims
7answer cells

A company that genuinely holds the credential but is not a beauty company.

Citation paradox

Does citing the official register mean the certification answer is right?

CertificationAssertionsCited register% citedDominant errorCause
EcoBeautyScore39127270%Conflation and stalenessPrimary source retrieved and misread
EWG Verified17411264%Rated confused with verifiedTwo EWG programmes conflated
CDP Triple A1411913%StalenessPrimary source not retrieved; trade press used instead
Science Based Targets1662012%ConflationParent group's targets attached to a consumer brand
Credential substitution

Is one certification being mistaken for another?

Seven claims name a real certification from the wrong scheme, across four different credential families. Each substitutes something real, adjacent and weaker for the specific credential that was asked about. That distinction is the entire reason a brand pays for the stronger scheme.

The answer saidThe brand actually holdsBrands
Leaping Bunny certifiedPETA Beauty Without BunniesUrban Decay, Wet n Wild
Science Based TargetsThe Climate PledgeSaie
EWG VerifiedAn EWG Skin Deep ratingEra Organics
CDP Triple AA single or double ACoty, Shiseido, Unilever
Monitoring heuristic

What can you trust from AI visibility tools — and what needs a different test?

Presence
Comparatively stable

A brand that appears in most runs will keep appearing. Trust it.

Rank position
Close to noise

A tool reporting a fixed AI rank is reporting something that will not survive being measured again. Ignore it.

Sentiment and comparison
In between

Directionally useful, individually volatile. Trend them; never act on a single reading.

Accuracy
A different question entirely

Not how often you appear or where, but whether what is said is true. It needs a source of truth outside the AI, which most brand claims do not have.

Business Implications

What should brands do when AI gets certification claims wrong?

01

A credential you earned can outlive its own expiry

Kao's CDP Triple A was real for five consecutive years and Symrise's made it one of only 14 companies worldwide in 2021. Both are reported in 2026 as current.

Audience

Any brand whose certification status has changed

First Step

A lapse, a rename or a merger is exactly when the answer layer is most likely to be wrong about you.

02

It can be attached to the wrong part of your own company

Beiersdorf holds the EcoBeautyScore membership. The answers credited Eucerin, a product brand. A group-level credential is being reported as a specific claim about a consumer product.

Audience

Groups with a portfolio of consumer brands

First Step

Test brand-level and group-level queries separately; they are different measurements.

03

It can outlive its own expiry

Kao's Triple A was real for five consecutive years and is being reported in 2026 as current, three scoring cycles after it last applied.

Audience

Any brand whose certification status has changed

First Step

A lapse or a rename is exactly when the answer layer is most likely to be wrong about you.

04

Citation share does not measure accuracy

Sixty-one of the 153 false-claim cells cited the certifying body's own register. Retrieval succeeded and the answer was still wrong.

Audience

Anyone buying AI visibility monitoring

First Step

Ask what a citation metric actually evidences. It shows a source was consulted, not that it was read correctly.

Methodology

How was this study designed — and what does it not claim?

270
Responses in scope
18
Credential prompts
Runs per prompt
5
AI engines
1,993
Citations
3
Markets
3
Languages
6
Credentials verified
Prompt categories verified
B Corp3 prompts
Leaping Bunny3 prompts
EWG Verified3 prompts
EcoBeautyScore3 prompts
CDP Triple A3 prompts
Science Based Targets3 prompts
Study design
  • Seven credential prompts per market across three markets, five engines, three runs each
  • Non-English prompts written natively in French and Spanish rather than translated
  • The cluster is complete and balanced at 21 responses per market-engine cell, with none missing
  • This report covers all six credentials with a public register, comprising 270 of the 315 responses
Ground truth
  • Every claim checked against the certifying body's own published register as it stood on 17 March 2026, the date the answers were generated
  • CDP 2025 A List, published 8 January 2026
  • EcoBeautyScore Association member list, published May 2025
  • EWG Verified directory and Skin Deep database, distinguishing the paid certification from the ratings database
  • Brand self-descriptions were not accepted as evidence of holding a credential
  • SBTi Companies Taking Action dashboard, a comprehensive public register, so absence from it is a meaningful negative
  • B Lab public directory for B Corp
  • Leaping Bunny CCIC (US) and Cruelty Free International (Europe) registers, which are separate programmes

Responses captured 17 March 2026; registers verified 4 August 2026

Exclusions
  • Clarins and myBlend are not published. One reading of the register puts them in the parent-only category since only Groupe Clarins has a B Lab profile, but Clarins' own certification announcement states the certificate covers both brands and the group. Sunday Riley and REIMI Botanicals could not be verified and are excluded.
  • 'Verified carbon commitments' names no scheme and has no register, so no answer to it can be resolved true or false.
  • Brands named only as portfolio context were removed. Eucerin and La Prairie appear in CDP answers as 'Beiersdorf: maker of Nivea, Eucerin, La Prairie', which names Beiersdorf as the holder and the others as description. Neither was claimed to hold the rating.
  • L'Oreal Group entities are excluded from named findings and included in aggregate counts.
  • Two B Corp claims were spot-checked and both were correct. They are reported as context and are not counted in the findings, because the remaining 41 B Corp claims are unverified.
Limits
  • This reports what was observed, not a rate. Three runs establishes that a claim was made and recurred; it cannot estimate how often it would recur.
  • A single execution date, 17 March 2026.
  • False denial was not tested. Establishing whether certified brands are being omitted requires checking against a complete register rather than a list of named brands.
  • The EcoBeautyScore member list is dated May 2025, ten months before the study, and the Association continuously accepts members. The eight lapsed-membership claims are the weakest in the set and are flagged for confirmation at right of reply.
  • CDP's press release of 8 January 2026 announced 23 Triple A companies; CDP's live scores page now shows 27. The four additions are post-appeal resolutions whose identity and date are not published. Beiersdorf, Shiseido and Unilever are reported as not holding Triple A on the basis of the January list and could be among the four. Each has been asked directly.
Statistical posture

Counts and ordinal persistence only. No rates are projected from this sample. At three runs per prompt, existence and recurrence are reported as observed facts, never as estimated frequencies.

Read your brand the way this report read the category

An AI Brand Representation Report measures one brand across all seven representation dimensions, in one market, with every finding traced to the prompts and responses behind it. AIVO does not sell the remediation.

Browse all industry research →

More AIVO Research

View all industry research →