SEEK
SEEK runs employment marketplaces in Australia, New Zealand and Southeast Asia (SEEK, Jobstreet and Jobsdb), and its seekAi employer tools label and rank applicants as 'high-fit' using a model trained on millions of past hirer shortlisting decisions, target job ads at predicted high-fit candidates, and run AI voice or text pre-screens, recommended shortlists and reference checks through a feature called Assist.
HireAIScore rates SEEK 54 out of 100 (grade F, Substantial gaps) under rubric v1.0, last reviewed September 25, 2026. SEEK ranks 33 of 160 vendors rated in Screening AI, against a category average of 46. That score is drawn from 30 evidence items across 7 rubric criteria for SEEK — 23 with a cited source, 7 recording that nothing was located.
§ 01 - Company facts
- Legal name
- SEEK Limited
- Founded
- 1997
- Headquarters
- Australia · AU
- Pricing tier
- Mid
- Market side
- Employer-side (AEDT)
- Categories
- ScreeningSourcing
- Website
- seek.com.au
These are company details, not findings. They are editable by a verified representative and carry no weight in the score, which is built only from the cited evidence below.
§ 02 - Score breakdown
§ Score breakdown
Category scoring
Weighted contribution shown to the right of each bar.
- 01
Article 11 Technical Documentation
Weight 20%70
+14.0 · category avg 49
- 02
Bias Audit Transparency
Weight 18%40
+7.2 · category avg 41
- 03
FRIA Support
Weight 15%32
+4.8 · category avg 31
- 04
Data Governance Disclosure
Weight 15%60
+9.0 · category avg 53
- 05
Human Oversight Design
Weight 12%62
+7.4 · category avg 56
- 06
Post-Market Monitoring
Weight 12%50
+6.0 · category avg 40
- 07
Customer Documentation
Weight 8%64
+5.1 · category avg 58
Category avg is the mean raw score on that criterion across the 160 Screening AI vendors in scope of this rubric, this one included.
§ 03 - Strongest · weakest
Strongest category
Article 11 Technical Documentation
Raw score 70 · contributes 14.0 to total.
70 against a 49 category average
Weakest category
FRIA Support
Raw score 32 · contributes 4.8 to total.
32 against a 31 category average
§ 04 - Cited evidence
Download diligence record→§ Evidence
Cited per category
Every score is backed by at least one cited piece of evidence.
Evidence ledger
- Items
- 30
- Documentation
- 18
- Audit report
- 2
- Public statement
- 3
- Absence
- 7
- With a source URL
- 23 of 30
- Source hostnames
- 8
- Fewest items
- 2
- FRIA Support
These figures measure how thoroughly SEEK was reviewed, not how SEEK performed — an absence row, recording that nothing was located, is counted like any other item.
Article 11 Technical Documentation
5 items70
- Audit reportCaptured Sep 25, 2026
SEEK's Trust Centre (hosted on Vanta) publicly shows a view-only ISO/IEC 42001:2023 certificate from Sustainable Certification (no. 2026-15271, certified 21 June 2026, expires 2 June 2029), which covers the procurement, design, development, operation, maintenance and retirement of AI services in the SEEK Marketplace's candidate and hirer services and certifies SEEK as an AI Producer.
- DocumentationCaptured Sep 25, 2026
SEEK's FY2026 Impact Report says it achieved ISO 42001 certification in FY2026, completed Responsible AI reviews of 94% of high-priority AI services (including matching and hirer-facing candidate discovery tools), and created an executive Responsible AI Forum.
https://au.seek.com/content/media/SEK-2026-Impact-Report.pdf
- DocumentationCaptured Sep 25, 2026
SEEK's Responsible AI page sets out four principles (Fair Opportunity, Transparent, Safe, Accountable) and a lifecycle governance framework informed by ISO 42001, the NIST AI RMF and Singapore's Model AI Governance Framework, with a detailed FAQ.
- DocumentationCaptured Sep 25, 2026
An employer help article explains that the high-fit label comes from AI that analyses millions of real-world shortlisting decisions together with profile, CV and job-ad data, and that 'Sort by high-fit' gives each application a high-fit score and ranks applicants in descending order.
https://help.au.employer.seek.com/s/article/All-You-Need-to-Know-About-Candidate-Management-ANZ
- AbsenceCaptured Sep 25, 2026
No model card, system card, Annex IV-style technical file or published accuracy and limitations statement was found for the seekAi high-fit model or Assist's ranking after searching the Responsible AI page, the Trust Centre's resources and controls, the employer product pages, the employer help centre, and the FY2026 Impact and Annual Reports.
Bias Audit Transparency
3 items40
- Public statementCaptured Sep 25, 2026
SEEK says it runs fairness audits with an internal RAI Assessment Tool that derives employment-specific fairness metrics and that 'external audits have been conducted', but it names no auditor, date, audited system or result.
- DocumentationCaptured Sep 25, 2026
The FY2026 Impact Report reports Responsible AI reviews of 94% of high-priority services but publishes no fairness metrics or findings from them, and its independent internal audit covered SEEK's fair-hiring and trust-and-safety controls rather than the ranking models.
https://au.seek.com/content/media/SEK-2026-Impact-Report.pdf
- AbsenceCaptured Sep 25, 2026
No NYC LL 144 bias-audit summary, impact-ratio table or independent disparate-impact audit of the high-fit/shortlisting model was found on SEEK's Responsible AI page, its Trust Centre, its sustainability and impact reports, its employer help centre, or the Warden AI trust dashboard (the SEEK path there returns 404).
FRIA Support
2 items32
- AbsenceCaptured Sep 25, 2026
No Fundamental Rights Impact Assessment template, EU AI Act Article 26/27 deployer guidance or AEDT notice kit was found on SEEK's Responsible AI page, its Trust Centre, its employer help centre or its product pages.
- DocumentationCaptured Sep 25, 2026
SEEK's Trust Centre offers a due-diligence package (public ISO 27001 and ISO 42001 certificates, control listings and governance policies), but its cyber security controls overview and 'Security & Privacy Questions' hirer guide are available only on request, and none of it addresses AI deployer obligations.
Data Governance Disclosure
5 items60
- DocumentationCaptured Sep 25, 2026
SEEK's Privacy Policy (last updated 14 August 2026) discloses that AI and machine learning use candidates' platform activity and application outcomes (shortlisted, interviewed, hired or rejected, including statuses from recruitment-software partners) to assess suitability against other candidates, and that AI pre-screen voice recordings are collected with consent, then transcribed and summarised.
- DocumentationCaptured Sep 25, 2026
SEEK's Selection Signals help article explains that applicant status updates flow from hirers' recruitment systems to SEEK so it can learn from their hiring activity and sharpen what high-fit looks like for each business.
https://help.au.employer.seek.com/s/article/What-are-Selection-Signals
- Audit reportCaptured Sep 25, 2026
SEEK's Trust Centre publicly displays its ISO/IEC 27001:2022 and ISO/IEC 42001:2023 certificates (2026) and lists its data and privacy controls, while a detailed cyber security controls overview requires an access request.
- DocumentationCaptured Sep 25, 2026
SEEK's candidate help article on the 'strong applicant' badge says the prediction uses the job ad, the employer's past behaviour and the candidate's default resume, and that SEEK does not use sensitive or personally identifiable information to generate the badge.
https://help.au.seek.com/s/article/What-is-the-You-might-be-a-strong-applicant-badge
- AbsenceCaptured Sep 25, 2026
No equivalent list of excluded features (such as names, age, gender, photos or location proxies) was found for the employer-side high-fit model or Assist's ranking after searching the Responsible AI page, the privacy policy, the Trust Centre and the employer help centre.
Human Oversight Design
6 items62
- DocumentationCaptured Sep 25, 2026
Each high-fit applicant in Candidate Management comes with an AI-generated summary of why they match and details of how their skills align, and hirers can sort by matching answers, star rating or newest application instead of high-fit.
https://help.au.employer.seek.com/s/article/All-You-Need-to-Know-About-Candidate-Management-ANZ
- DocumentationCaptured Sep 25, 2026
SEEK's AI pre-screens are optional for candidates, use a fixed and published question set, and let candidates review and edit their answers or voice summary before submitting, and SEEK states they 'do not decide, score or recommend whether a candidate should progress'.
https://help.au.employer.seek.com/s/article/What-is-an-AI-pre-screen
- Public statementCaptured Sep 25, 2026
SEEK states that its services 'do not make any final decisions to shortlist, interview or hire a candidate', that all applications remain accessible to hirers, and that users can contest AI-driven decisions through customer service.
- DocumentationCaptured Sep 25, 2026
Assist's next-best-action emails include a 'Recommended shortlist' that explains why applicants may fit, and SEEK states that Assist 'does not decide who should be hired'.
- DocumentationCaptured Sep 25, 2026
A 'must-have' screening filter is on by default, and hirers can choose to auto-move applicants who don't meet it to 'Not Suitable', after which SEEK emails those candidates within 48 hours, an automated-rejection path with no documented review step.
https://help.au.employer.seek.com/s/article/All-You-Need-to-Know-About-Candidate-Management-ANZ
- AbsenceCaptured Sep 25, 2026
No documented per-jurisdiction controls, AI audit logs or option to switch off high-fit labelling or sorting was found in SEEK's employer help centre or product pages.
Post-Market Monitoring
4 items50
- DocumentationCaptured Sep 25, 2026
SEEK runs a public Atlassian Statuspage for the SEEK API with 14 components (including Selection Signals, SEEK Profiles and Ad Performance), email, SMS, Slack and RSS subscriptions, and an active incident history (its 50 most recent incidents run from August 2024 to September 2026), although no incident has a postmortem.
- Public statementCaptured Sep 25, 2026
SEEK runs a public Bugcrowd bug bounty engagement for reporting security vulnerabilities, listed as in progress since November 2016.
- DocumentationCaptured Sep 25, 2026
SEEK reports each year how many of its high-priority AI services received Responsible AI reviews (94% in both FY2026 and FY2025) and commits to reviewing the rest in the following year.
https://au.seek.com/content/media/SEK-2026-Impact-Report.pdf
- AbsenceCaptured Sep 25, 2026
No public AI model-update changelog, AI-specific incident or adverse-impact reporting channel beyond general customer service, or public fairness or drift monitoring dashboard was found on the Responsible AI page, the Trust Centre, the status page or the help centres.
Customer Documentation
5 items64
- DocumentationCaptured Sep 25, 2026
SEEK's 'AI at SEEK' product page describes how Assist scores and ranks applicants and explains why each fits, along with AI voice interviews, automated reference checks and Personalised Targeting.
- DocumentationCaptured Sep 25, 2026
SEEK's employer help centre documents the AI pre-screen in detail, including the exact questions asked, allowances by ad type and the 7-day expiry, alongside articles on Assist, Selection Signals, Personalised Targeting and Candidate Matches.
https://help.au.employer.seek.com/s/article/What-is-an-AI-pre-screen
- DocumentationCaptured Sep 25, 2026
Jobstreet's candidate help article explains the AI pre-screen, including that voice recordings are not shared with hirers, that transcripts are kept only for a limited period for quality checks, and that Jobstreet does not decide who progresses.
https://help.id.jobstreet.com/s/article/What-is-an-AI-pre-screen-candidates-ID
- DocumentationCaptured Sep 25, 2026
SEEK's Responsible AI page, mirrored on the Jobstreet and Jobsdb sites, hosts an FAQ covering consent, data use, bias mitigation, contestability, opt-outs and third-party AI procurement.
- AbsenceCaptured Sep 25, 2026
No public DPA, NYC LL 144 or EU AI Act deployer guidance, or AI-specific compliance pack for hirers was found, and the Trust Centre's 'Security & Privacy Questions' hirer guide is available only on request.
§ 05 - Editorial notes
Company overview
SEEK Limited (ASX: SEK) was founded in Melbourne in 1997 and describes itself as Australia's #1 job site. It runs SEEK in Australia and New Zealand, Jobstreet in Malaysia, Singapore, the Philippines and Indonesia, and Jobsdb in Hong Kong and Thailand. One employer platform serves all eight markets (talent.seek.com.au now redirects to au.employer.seek.com). SEEK also wholly owns JobAdder, a recruitment-agency ATS and CRM that is scored separately. Its employer-side AI is branded seekAi. SEEK says seekAi combines resume and profile models with a model trained on millions of real hirer shortlisting decisions to predict how likely each applicant is to be shortlisted. It then labels applicants 'high-fit', attaches an AI-generated rationale, and sorts them by a high-fit score. Personalised Targeting learns from each hirer's connects, shortlists and hires, including status updates sent from applicant tracking systems, which SEEK calls Selection Signals. Assist adds recommended-shortlist emails, optional AI voice or text pre-screens and automated reference checks. Candidate Matches surfaces up to 100 AI-matched profiles for each live ad. SEEK earns its revenue from tiered job-ad products (Lite in Asia, then Basic, Advanced and Premium), which puts it in the mid price band.
Regulatory exposure
All of SEEK's marketplaces are in APAC. Its exposure to NYC Local Law 144, Illinois HB 3773, Colorado SB 26-189 and California's FEHA rules on automated-decision systems is therefore limited to hirers who use SEEK for roles in those places, and SEEK has published no LL 144 bias audit. In function, though, the high-fit score and Assist's ranking score and rank applicants to substantially assist shortlisting. That is the kind of tool LL 144 treats as an AEDT, and the EU AI Act lists such tools as high-risk under Annex III point 4. Those EU obligations now apply from 2 December 2027. With no EU marketplace, SEEK's EU exposure comes only through multinational hirers. The central compliance question is its training on historical hirer shortlisting decisions and per-hirer feedback loops, which can reproduce past bias. SEEK's documentation says the AI pre-screen only transcribes and summarises answers, with no scoring and no emotion analysis, so on that evidence the Article 5(1)(f) prohibition on emotion recognition is not engaged. The pre-screen is voice or text only, so Illinois' AI Video Interview Act would not be triggered. SEEK aligns its framework with Australia's AI Ethics Principles and the Voluntary AI Safety Standard and with Singapore's Model AI Governance Framework, and it joined Singapore's AI Verify Foundation in FY2026. Its FY2026 Annual Report lists biased or unfair AI outcomes and tightening regulation among its principal risks.
Path to a higher score
SEEK's governance is ahead of its model-level transparency. The largest gain would come from an independent bias audit of the high-fit/shortlisting model and of Assist's ranking. It should be published and repeated each year, with selection-rate impact ratios by sex, age band and, where lawful, ethnicity, plus a named auditor and date. That matters most because the model learns from historical hirer shortlisting and from per-hirer signals (Personalised Targeting and Selection Signals). Next, SEEK should publish a system card for high-fit covering inputs, performance, limitations and excluded features such as names, photos, age and location proxies, extending the exclusion statement it already makes for the candidate-side strong-applicant badge. It should also give multinational hirers deployer guidance (EU AI Act Articles 26 and 27, LL 144 notice and audit duties) and a public DPA. An AI model-update changelog and a dedicated channel for reporting AI issues would help too. Finally, it should document hirer controls: audit logs, a way to switch off high-fit labelling, and a review step before must-have filters auto-reject applicants.
§ Regulatory frame
What applies to screening ai.
Treated as high-risk under Annex III §4 when the screening output is decisive. EU AI Act Article 13 transparency obligations apply; deployers must give candidates a way to contest.
§ Compare
Build any comparison→SEEK against its nearest-scoring peers.
In Screening AI.
§ Others rated in Screening AI
All screening vendors→Ranked 33 of 160 by weighted total under rubric v1.0. The ordering is arithmetic on the rubric and carries no view on which tool suits a given hiring process.
- 30Checkr54F
- 31Bullhorn54F
- 32Codility54F
- 33SEEKThis profile54F
- 34Spark Hire53F
- 35Criteria53F
Conflicts of interest
No vendor pays for placement, scoring, or removal. Casework - the consulting firm that operates this directory - provides paid services to some vendors. Any active or recent (within 24 months) commercial relationship is disclosed on the affected vendor profile and the review is reassigned to an independent reviewer. See the full policy on About.
Casework has no commercial relationship with this vendor.