8/8 publish an explicit boundary or control around candidate or employer AI use in at least one part of recruitment.
Where AI assistance ends
The Who Gets In? AI and Candidate-Owned Performance Tracker compares what eight reviewed institutions explicitly publish about candidate preparation, protected assessments and interviews, employer use of AI and retained human authority. It uses current official sources and keeps the denominator beside every finding.
What does the reviewed evidence establish?
7/8 explicitly protect at least one assessment or interview context through an AI restriction or candidate-owned-performance requirement in the reviewed sources; BCG is not coded into this numerator because its reviewed guidance does not establish a stage-specific live-assessment prohibition.
7/8 explicitly permit or encourage some form of AI-supported preparation or application assistance in the reviewed source set; PwC UK was not coded as permitting preparation because the reviewed source addresses assessment responses rather than general preparation.
Human involvement is explicit at all 8 institutions, but the control differs materially across recruiter review, people making decisions, human assessment review, application reading, live assessor interviews and assessment centres/Super Days.
How does each institution draw the boundary?
| Institution | Preparation/application | Protected stage | Employer AI | Human authority | Primary evidence |
|---|---|---|---|---|---|
| McKinsey & CompanyGlobal | encouraged for responsible preparation such as resume polishing and interview practice | prohibited for real-time interview answers and assessments unless specifically permitted | AI-assisted interview exists as a voluntary route/policy architecture | AI-assisted interview assessments are reviewed by a human recruiter before a rejection decision | Official source |
| UBSGlobal | permitted for application documents where output genuinely reflects the candidate | assessment tasks and recorded/virtual video responses must be completed independently without AI-generated responses | not inferred from candidate-use policy | UBS states people make decisions on submissions | Official source |
| PwCUnited Kingdom | not coded as permitted from the reviewed assessment-policy source | AI or third-party generation, editing, rewriting, summarising or coaching of assessment responses is prohibited unless authorised | PwC states it does not use AI to assess Early Careers candidate performance | on-demand video answers are reviewed by a human assessor | Official source |
| Clifford ChanceUnited Kingdom | AI tools may support application preparation but overreliance is discouraged and authentic voice is emphasised | AI is not permitted in the Watson Glaser test | The Challenge separately uses AI tools inside a trainee-style simulation; this is not treated as general recruitment AI | the firm states its team reads every application | Official source |
| DeloitteUnited Kingdom | responsible AI use is permitted for selection-process preparation, final-stage topic research and interview practice, while candidate interview responses must remain based on their own experiences | AI is prohibited in named online assessments and in both in-person and virtual interviews; those stages must be completed independently | not inferred from candidate-use guidance | the Early Careers process ends with live assessor discussion and/or a skills and motivation interview depending on programme | Official source |
| HSBCGlobal | GenAI may be used to research HSBC and roles, structure a CV or application, prepare for interviews and organise or draft ideas where the final content remains accurate and genuinely the candidate's own | GenAI must not generate responses during live or recorded interviews; candidates must not submit work they cannot personally explain, validate or replicate | not inferred from candidate-use guidance | the student and graduate process publishes an assessment centre/Super Day after the online assessment for applicable programmes | Official source |
| BCGGlobal | GenAI may support brainstorming, structuring resumes and responses, source-finding, clarity checks and interview practice, while the candidate remains responsible for authentic content and validation | the reviewed candidate guidance requires authentic candidate-owned application content but does not establish a blanket prohibition for live interviews or assessments; those stages remain not yet evidenced for this coding field | BCG describes limited AI support including job recommendations and drafting interview questions, while stating AI is never used to make hiring decisions | BCG states applications are reviewed by experienced recruiters and hiring decisions are made by experienced recruiters and hiring managers | Official source |
| A&O ShearmanUnited Kingdom | AI may complement candidate skills and help refine CVs, applications and written responses, but should not substitute the candidate's voice or misrepresent skills or experience | the candidate charter states that applications and assessments should reflect the candidate's own thinking, experience and abilities | A&O Shearman's London early-careers FAQ states it does not use AI at any point during recruitment | the same FAQ states all assessments and decisions are made by the firm's team; the London training-contract route also publishes two one-to-one interviews after the online experience | Official source |
What does this not prove?
- N=8 is a defined evidence subset, not a prevalence estimate for elite employers.
- Policies are institution-, market-, programme- and stage-specific and may change.
- Absence of a coded permission or prohibition is not evidence that no such policy exists.
- A general authenticity requirement is not automatically coded as a prohibition on AI during every assessment stage.
Use and cite the tracker
Preferred wording: According to the Who Gets In? AI and Candidate-Owned Performance Tracker, published by Elite Careers Strategy, …
Download the governed tracker as JSON · Read the bounded answer · Methodology · Journalist citation desk