LinkedIn Recruiter
LinkedIn Recruiter is LinkedIn's enterprise sourcing and candidate-management platform, whose Hiring Assistant add-on and AI-Assisted Search use AI to turn intake notes into qualifications, source and rank candidates, evaluate applicants against recruiter-set qualifications, and draft or send outreach.
HireAIScore rates LinkedIn Recruiter 57 out of 100 (grade F, Substantial gaps) under rubric v1.0, last reviewed September 25, 2026. LinkedIn Recruiter ranks 10 of 86 vendors rated in Candidate sourcing AI, against a category average of 46. That score is drawn from 39 evidence items across 7 rubric criteria for LinkedIn Recruiter — 34 with a cited source, 5 recording that nothing was located.
§ 01 - Company facts
- Legal name
- LinkedIn Corporation
- Founded
- 2002
- Headquarters
- United States · US
- Pricing tier
- Enterprise
- Market side
- Employer-side (AEDT)
- Categories
- SourcingScreening
- Website
- business.linkedin.com
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%67
+13.4 · category avg 48
- 02
Bias Audit Transparency
Weight 18%50
+9.0 · category avg 40
- 03
FRIA Support
Weight 15%38
+5.7 · category avg 31
- 04
Data Governance Disclosure
Weight 15%62
+9.3 · category avg 53
- 05
Human Oversight Design
Weight 12%66
+7.9 · category avg 56
- 06
Post-Market Monitoring
Weight 12%50
+6.0 · category avg 40
- 07
Customer Documentation
Weight 8%68
+5.4 · category avg 57
Category avg is the mean raw score on that criterion across the 86 Candidate sourcing AI vendors in scope of this rubric, this one included.
§ 03 - Strongest · weakest
Strongest category
Customer Documentation
Raw score 68 · contributes 5.4 to total.
68 against a 57 category average
Weakest category
FRIA Support
Raw score 38 · contributes 5.7 to total.
38 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
- 39
- Documentation
- 30
- Audit report
- 1
- Public statement
- 3
- Absence
- 5
- With a source URL
- 34 of 39
- Source hostnames
- 8
- Fewest items
- 4
- FRIA Support
These figures measure how thoroughly LinkedIn Recruiter was reviewed, not how LinkedIn Recruiter performed — an absence row, recording that nothing was located, is counted like any other item.
Article 11 Technical Documentation
5 items67
- DocumentationCaptured Sep 25, 2026
The 19-page Hiring Assistant Data Security & Compliance whitepaper (file dated March 2026) covers LinkedIn's design review, risk assessments and DPIAs, training and runtime data, an 'AI explainability' walkthrough, and which model runs each part: OpenAI models via Azure for intake and search, and an in-house LLM that marks each qualification as matched, partly matched or not matched and cites the source.
https://business.linkedin.com/content/dam/lem/business/en/hire/ai-transparency/liha-white-paper.pdf
- DocumentationCaptured Sep 25, 2026
A public data-flow diagram shows a supervisor agent directing intake, sourcing, evaluation and outreach sub-agents, with inference on Microsoft Azure and an in-house open-source model fine-tuned on LinkedIn Economic Graph data, and states that customer data is not used for fine-tuning.
- DocumentationCaptured Sep 25, 2026
An October 2025 engineering post describes the plan-and-execute supervisor design with seven sub-agents, custom LLMs that evaluate qualifications and show recruiters the supporting evidence, quality benchmarks run before release, and Responsible AI safety checks on every qualification.
https://www.linkedin.com/blog/engineering/ai/how-we-engineered-linkedins-hiring-assistant
- DocumentationCaptured Sep 25, 2026
The Trust Center's 'AI - LinkedIn Talent Solutions' section offers the Hiring Assistant AI data compliance whitepaper, whitepapers for AI-Assisted Messages and AI-Assisted Search & Projects, and a gated Hiring Assistant AI flow diagram, and its list of certifications includes no ISO/IEC 42001.
- AbsenceCaptured Sep 25, 2026
No ISO/IEC 42001 certification, no model card with accuracy, error-rate or validation figures for the evaluation LLM, and no EU AI Act-style technical file or instructions for use turned up on the AI transparency hub, the Recruiter help center, the Trust Center or security.linkedin.com.
Bias Audit Transparency
5 items50
- Audit reportCaptured Sep 25, 2026
An independent audit by university researchers from USC, Princeton and Stony Brook (Behzad, Devic, Sharan, Korolova, Kempe; posted 13 Nov 2025, free to download) tested LinkedIn Recruiter's Talent Search ranking through Recruiter Lite without LinkedIn's knowledge. It found minority groups under-represented in early ranks across many queries and uneven ranking stability between groups, and concluded that LinkedIn's apparent demographic-aware post-processing works less well at the very top ranks.
- DocumentationCaptured Sep 25, 2026
LinkedIn's own KDD 2019 paper reports that fairness-aware re-ranking in Talent Search nearly tripled the share of queries with representative results in A/B tests and was rolled out to 100% of LinkedIn Recruiter users worldwide; this is a self-assessment, not an independent audit.
- Public statementCaptured Sep 25, 2026
LinkedIn's August 2026 guide 'Responsible AI in Hiring' says LinkedIn checks Hiring Assistant models for bias before deployment, audits them regularly afterwards, and retrains or pulls back models when problems appear, but it names no auditor or date and publishes no results.
- DocumentationCaptured Sep 25, 2026
A May 2024 engineering post describes LinkedIn's privacy-preserving method for estimating race and ethnicity so it can measure 'Equal Treatment' in its U.S. algorithms, but gives no results for Recruiter or Hiring Assistant.
- AbsenceCaptured Sep 25, 2026
No NYC Local Law 144 bias-audit summary or impact ratios for Recruiter, Hiring Assistant or AI applicant ranking were found on the AI transparency hub, the Recruiter help center compliance FAQs, the Trust Center's AI and Reports sections, or Warden AI's dashboard, which returns 404 for LinkedIn; targeted searches also found no audit by BABL AI, DCI, ORCAA, Holistic AI or Credo AI.
FRIA Support
4 items38
- DocumentationCaptured Sep 25, 2026
The Talent Solutions AI Compliance FAQ says LinkedIn carries out risk assessments and DPIAs for its generative-AI products and will give customers product information for their own compliance assessments, including DPIAs.
- DocumentationCaptured Sep 25, 2026
LinkedIn's guide for employers gives a five-part evaluation framework (Control, Clarity, Reviewability, Readiness, Trust) and cites the EU AI Act's emphasis on human oversight, documentation, AI literacy and governance, but lists no deployer obligations, FRIA template or Article 26/27 steps.
- Public statementCaptured Sep 25, 2026
The whitepaper mentions the AI Act only once, in a general statement that LinkedIn respects laws 'including EU legislation like the AI Act' and is watching emerging requirements closely.
https://business.linkedin.com/content/dam/lem/business/en/hire/ai-transparency/liha-white-paper.pdf
- AbsenceCaptured Sep 25, 2026
No Fundamental Rights Impact Assessment template, Article 26/27 deployer guidance or EU AI Act risk classification of Hiring Assistant was found on the AI transparency hub, the Hiring Assistant help pages or the Trust Center's 'AI - LinkedIn Talent Solutions' section.
Data Governance Disclosure
5 items62
- DocumentationCaptured Sep 25, 2026
The Hiring Assistant FAQ lists the profile fields, resumes, screening answers and 'World Knowledge' used to source and evaluate candidates, excludes mutual connections and contact instructions, and blocks qualifications or prompts about gender, race, class, disability, marital or maternity status, sex, religion, sexual orientation, socio-economic factors and age.
- DocumentationCaptured Sep 25, 2026
The whitepaper says training data comes from LinkedIn's Economic Graph (profiles, experience, skills, education, certifications, location, job postings); customer inputs count as Customer Personal Data under the DPA and may be used to improve search or train future in-house models; and LinkedIn data is not used to train third-party models, including OpenAI's.
https://business.linkedin.com/content/dam/lem/business/en/hire/ai-transparency/liha-white-paper.pdf
- DocumentationCaptured Sep 25, 2026
The AI Search and Projects FAQ bans race, gender, nationality, sexuality and age filters and explains that machine-learning models rank results by how closely candidates match the search and how likely they are to respond.
- DocumentationCaptured Sep 25, 2026
LinkedIn's Trust and Compliance page lists ISO/IEC 27001 and ISO/IEC 27018 certifications and a SOC 2 report covering Talent Solutions, plus ISO 22301 for the platform.
- DocumentationCaptured Sep 25, 2026
The LinkedIn Trust Center lists a SOC 2 Type II report audited by Baker Tilly (latest period ending 30 September 2025, access-gated), ISO 27001 with a Statement of Applicability, ISO 27018, ISO 22301 and EU-US DPF, but no ISO/IEC 42001.
Human Oversight Design
6 items66
- DocumentationCaptured Sep 25, 2026
The help page on reviewing applicants explains that sorting by Relevance puts Hiring Assistant's top applicants first (Screening Requirements, Newest and Alphabetical are alternatives), that recruiters can open an evaluation against each qualification showing whether the evidence came from the profile or the resume, and that evaluation can be paused, stopped or capped from the Activity log.
- DocumentationCaptured Sep 25, 2026
AI-Assisted Search shows an AI summary of fit and a High (70%+), Medium (30%+) or Low (<30%) relevance label that recruiters can click for examples, and recruiters can make keyword search their default instead.
- DocumentationCaptured Sep 25, 2026
The Hiring Assistant FAQ says the tool will not screen out applicants or make hiring decisions, sends outreach and follow-ups automatically only if a recruiter turns instant outreach on, cannot turn instant outreach on for projects located in the EU, and labels messages 'Sent by [recruiter's first name]'s Hiring Assistant'.
- DocumentationCaptured Sep 25, 2026
Admins can turn off Hiring Assistant's applicant evaluation for the whole contract, but it is on by default.
- DocumentationCaptured Sep 25, 2026
The learning-and-feedback page explains that Hiring Assistant adapts to explicit feedback and to implicit signals (searches, candidates added to the pipeline, messages sent), and that suggested changes to qualifications take effect only after the recruiter reviews and accepts them.
- DocumentationCaptured Sep 25, 2026
LinkedIn's guide describes in-context explanations of why a candidate may match and an Activity log that lets customers look back over Hiring Assistant's actions for traceability.
Post-Market Monitoring
7 items50
- AbsenceCaptured Sep 25, 2026
No public AI incident channel, Hiring Assistant model-update changelog or continuous fairness-monitoring dashboard was found; LinkedIn's hiring release pages announce new features without describing model changes.
- DocumentationCaptured Sep 25, 2026
LinkedIn's public Atlassian Statuspage tracks LinkedIn Recruiter, Jobs, Talent Hub and Talent Insights under Talent Solutions and keeps an incident history, but it has no Hiring Assistant or AI component.
- DocumentationCaptured Sep 25, 2026
The status page's incident feed shows Recruiter outages on 15 March 2024 and 9 October 2024 and a Talent Solutions-wide incident on 6 March 2024, each recorded only as a short 'investigating' and 'resolved' note with no root-cause analysis.
- DocumentationCaptured Sep 25, 2026
A separate LinkedIn API status page records Talent Solutions integration incidents, such as blocked Apply Connect webhooks (22 May 2026) and a failed Talent Solutions onboarding widget (18 June 2026).
- DocumentationCaptured Sep 25, 2026
LinkedIn runs a public HackerOne bug bounty for reporting security vulnerabilities.
- DocumentationCaptured Sep 25, 2026
Customers get a Hiring Assistant Report on usage and performance across sourcing, applicant evaluations, top-applicant ratings, messaging and prescreening, with two years of history and XLS export, but it includes no demographic or adverse-impact metrics.
- Public statementCaptured Sep 25, 2026
LinkedIn says it continuously monitors Hiring Assistant's stability, latency, output-quality signals, failure rates and user feedback, and regularly evaluates model quality, but it publishes none of the results.
Customer Documentation
7 items68
- DocumentationCaptured Sep 25, 2026
The AI Transparency hub for LinkedIn's hiring and learning products links the Hiring Assistant whitepaper, data-flow diagrams, general and compliance FAQs, the AI-Assisted Search whitepaper, ATS data-security guides and the Trust Center, and sets out LinkedIn's Responsible AI principles.
- DocumentationCaptured Sep 25, 2026
A transparency page for Hiring Assistant, AI-Assisted Search, AI-Assisted Messages and RSC+ covers data use, human involvement, excluded sensitive data, bias reviews and how candidates are ranked.
- DocumentationCaptured Sep 25, 2026
The Hiring Assistant product page describes sourcing, applicant evaluation, outreach and prescreening, and says the tool is sold as an add-on to LinkedIn Recruiter.
- DocumentationCaptured Sep 25, 2026
The Hiring Assistant in Recruiter FAQ (updated September 2026) pairs feature questions with a compliance section and links to how-to articles on admin settings, applicant review, prescreening, reporting and webinars.
- DocumentationCaptured Sep 25, 2026
LinkedIn's public Data Processing Agreement covers Talent/Hire products and includes Standard Contractual Clauses, a subprocessor list, a schedule of security measures and audit-report rights.
- DocumentationCaptured Sep 25, 2026
LinkedIn's 16-page guide 'Responsible AI in Hiring: What HR and TA Professionals Should Look For' (August 2026) gives employers an evaluation framework and explains how Hiring Assistant supports control, clarity, reviewability, readiness and trust.
- AbsenceCaptured Sep 25, 2026
No guidance for employers on NYC Local Law 144, Illinois HB 3773 or Colorado's AI law, and no candidate-notice template, was found on the AI transparency hub, the Hiring Assistant help pages, the Talent Solutions AI Compliance FAQ or the Trust Center.
§ 05 - Editorial notes
Company overview
LinkedIn Corporation was founded in 2002, is based in Sunnyvale, California, and has been a Microsoft subsidiary since December 2016. LinkedIn Recruiter is its enterprise sourcing and candidate-management product, priced by quote. This assessment covers the AI inside Recruiter: Hiring Assistant, AI-Assisted Search and AI applicant ranking. Hiring Assistant is an agent sold as an add-on. It was piloted with charter customers from October 2024, became generally available in English in late September 2025, and now runs in seven languages, including German and French. A supervisor agent directs sub-agents that turn intake notes into qualifications, run sourcing searches, and evaluate LinkedIn and ATS applicants one qualification at a time. They also list 'top applicants' first, draft or automatically send InMails, and run prescreening chats. Intake and search use OpenAI models through Azure, and candidate evaluation runs on LinkedIn's own LLMs.
Regulatory exposure
Hiring Assistant evaluates and ranks job applicants. That is the employment use the EU AI Act lists as high-risk in Annex III, and under the Digital Omnibus those obligations apply from 2 December 2027. LinkedIn's documents mention the AI Act only in a general line about respecting applicable law. They present the product as decision support that 'will not screen out applicants or make hiring decisions'. Even so, a New York City employer that relies on the top-applicant rating could be using an automated employment decision tool under Local Law 144. LinkedIn publishes no LL 144 bias audit, no candidate-notice template and no NYC setting. An independent academic audit of Recruiter's search ranking in November 2025 found minority groups under-represented in the top ranks. Illinois HB 3773 (in force since January 2026) and Colorado SB 26-189 (effective January 2027) add notice and documentation duties. The only location-based control LinkedIn documents turns off auto-outreach for projects in the EU.
Path to a higher score
The quickest gains are in bias transparency and support for deployers. LinkedIn should commission an independent LL 144-style audit of Hiring Assistant's applicant evaluation, its top-applicant rating and AI-Assisted Search ranking, and publish selection rates and impact ratios by sex, race/ethnicity and intersectional groups. It should also publish the results of the 'regular bias audits' it says it already runs. For the EU, it should state how Hiring Assistant is classified under Annex III and publish instructions for use, Article 26/27 deployer guidance and a FRIA template. For the US, it should provide NYC, Illinois and Colorado notice templates and a location setting that switches off evaluation for candidates in regulated places. Beyond that, a model card with accuracy and error rates for the evaluation model, ISO/IEC 42001 certification, a public model-update changelog, and adverse-impact metrics in the customer-facing Hiring Assistant Report would move it higher.
§ Regulatory frame
What applies to candidate sourcing ai.
Ranking outputs are inside Annex III §4 when they influence hiring decisions. GDPR Article 22 (automated decision-making) and the GDPR rights around profiling apply independently for EU candidates.
§ Compare
Build any comparison→LinkedIn Recruiter against its nearest-scoring peers.
In Candidate sourcing AI.
§ Others rated in Candidate sourcing AI
All sourcing vendors→Ranked 10 of 86 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.
- 07Fuel5058F
- 08Findem58F
- 09Pinpoint57F
- 10LinkedIn RecruiterThis profile57F
- 11Juicebox (PeopleGPT)57F
- 12Rippling56F
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.