Research by Elite Careers Strategy

How the world's most selective institutions choose people.

What selective employers and universities publish about how they choose people, read institution by institution and market by market, with every figure tied to a dated official source.

As featured in

  • Financial Times

Hassan Akram, who leads this research, has been featured in the Financial Times. Shown for context only. No affiliation, sponsorship, or endorsement is implied. The Financial Times did not review or commission this research.

31 of 67 employers checked publish a rule on candidates' AI use

Employers allow AI help on the written application but draw the line at interviews

Of 31 selective employers that publish a rule on candidates' use of AI, 25 allow AI help with the written application and none forbid it; 18 forbid AI in interviews and none allow it.

Located
37 of 67 publish an official statement on AI in selection
Application
25 of 31 allow AI help on the written application
Interviews
18 of 31 forbid AI in interviews; none allow it
Who decides
16 state that people, not AI, make or review the decision

Limit Does not show that the counts describe all selective employers (the list is purposive, not a sample). "Not located" never means "no policy".

AI Rules Tracker v2.1, observed 23 October 2026Lead finding AIR-2.0-F1Sector views: banks, law, consulting and Big Four

Look up an employerCite and download

ExhibitAI Rules Tracker v2.1

Candidate AI rules by selection stageNumber of employers whose official pages address candidates' own use of AI (n = 31 of 67 checked). Observed 23 October 2026.
  • Allowed
  • Limited or own-work only
  • Forbidden
  • Not stated
  1. Written application25 of 31 allowed6 of 31 not stated
  2. Online tests1 of 31 allowed4 of 31 limited or own-work only14 of 31 forbidden12 of 31 not stated
  3. Interviews3 of 31 limited or own-work only18 of 31 forbidden10 of 31 not stated
  4. Final-stage tasks1 of 31 allowed2 of 31 forbidden28 of 31 not stated

"Allowed" includes allowed with limits; "Forbidden" includes forbidden unless the employer authorises it. A purposive list, not a sample; published rules are not evidence of enforcement. Source: Hassan Akram, Who Gets In? AI Rules Tracker v2.1, observed 23 October 2026. CC BY 4.0. Chart for reuse (SVG)Figures as a table

Findings

What the evidence currently supports

4 further findings from smaller defined sets, each with its denominator and the claim it does not support, reviewed through 13 September 2026.

All findings, including the frozen report findings

  1. Finding 1

    Explicit AI rules draw a boundary around candidate-owned work

    8 of 8 institutions in the defined subset

    Limit Do not generalise 8/8, or the stage-specific 7/8 result, to all selective employers.

  2. Finding 2

    Human judgement remains visible after digital selection stages

    7 of 7 institutions in the defined subset

    Limit This describes the reviewed subset only; it is not a prevalence estimate for elite employers.

  3. Finding 3

    The same employer can run materially different selection architecture by market

    3 of 3 institutions in the defined subset

    Limit Not a claim that every multinational employer varies by geography.

  4. Finding 4

    A single BlackRock programme gives London and GCC candidates different application runways

    1 of 1 institutions in the defined subset

    Limit This is a programme-specific cross-market observation, not evidence that all BlackRock or elite-employer programmes use different regional deadlines.

Global Selection Atlas

How deep the evidence is in each of the five cities

The Global Selection Atlas is a depth product, not the research universe. For 8 priority institutions it records the state of the official evidence in each of 5 hubs. A dossier is promoted only where the evidence is deep enough for bounded cross-market comparison; every other institution in the register has an evidence record. "Not yet evidenced" never means no.

Markets and hub comparison

1387 dated facts (signals) from 494 official source records, across a 190-employer research census, 7 sectors and 5 priority market groups.

  • Employers in research census 190
  • Priority market groups 5
  • Selection question families 14
  • Employer Census denominators: see the research scope
Institutions by hub, with evidence state
InstitutionLondonNew YorkHong KongSingaporeDubai/GCCPage type
BlackRock5 verified of 5 hubs12 city observationsVerifiedVerifiedVerifiedVerifiedVerifiedDossier
Goldman Sachs5 verified of 5 hubs12 city observationsVerifiedVerifiedVerifiedVerifiedVerified. Scope note: Current official Goldman Sachs recruiting evidence establishes a named 2027 Dubai Investment Banking Capital Solutions Group Seasonal/Off Cycle Internship. The route is for undergraduate students, includes training and real responsibilities, and publishes a four-business/location application limit per recruiting year. Evidence is route-specific and is not represented as a universal Dubai or Goldman selection model.Dossier
JPMorganChase3 verified, 1 partial, 1 not yet evidenced of 5 hubs8 city observationsPartialVerified. Scope note: Current official 2027 Global Investment Banking Summer Associate evidence establishes New York as a programme location, with all listed groups available there, a nine-week programme, four days of orientation/training, MBA eligibility and a resume-based application review. This is the MBA Summer Associate route and is not silently treated as the undergraduate Summer Analyst route.VerifiedVerifiedNot yet evidencedEvidence record
Morgan Stanley4 verified, 1 partial of 5 hubs8 city observationsVerifiedPartialVerifiedVerified. Scope note: Current 2027 Investment Banking Industrial Placement / Summer Analyst route covers Hong Kong, Singapore and Seoul; Asia excluding India candidates may apply to up to three opportunities per recruitment year.Verified. Scope note: Current 2027 Dubai Investment Banking Off-Cycle Internship runs six months from January 2027; deadline 18 October 2026; rolling recruitment.Dossier
Clifford Chance5 verified of 5 hubs16 city observationsVerifiedVerified. Scope note: Current official New York/US evidence now covers application materials, first-semester-grade update timing, rolling review across US summer-law-clerk recruiting and a bounded cross-office interview route involving a New York attorney.VerifiedVerified. Scope note: Current candidate-facing 2027 Summer Internship page publishes the 29 January 2027 deadline, online application, Watson Glaser, associate/Graduate Recruitment interview, additional partner/counsel interview for Practice Training Contract applicants, and an internship-to-Practice-Training-Contract interview route.VerifiedDossier
Freshfields5 verified of 5 hubs15 city observationsVerifiedVerified. Scope note: Current official US law-student evidence establishes a ten-week New York summer-associate programme and publishes callback architecture plus practice-interest matching; US callback evidence is not assumed identical across every route or cycle.VerifiedVerified. Scope note: Current Singapore IAG Summer Internship 2027 evidence establishes the 1 October-31 December 2026 application window, three-to-six-week Singapore placement and required CV, cover letter, legal writing sample and transcripts.Verified. Scope note: Current Dubai/MENA evidence establishes office-specific screening, a Dubai Dispute Resolution Internship application-review-to-interview sequence, and a 2027 Emirati Internship with application, interview and offer stages.Dossier
HSBC5 partial of 5 hubs3 city observationsPartialPartialPartialPartialPartialEvidence record
UBS5 partial of 5 hubs6 city observationsPartialPartialPartialPartialPartialEvidence record

Show one hub

VerifiedVerified in the official sourcePartialPartial: evidence exists but is not programme-level for that hubNot yet evidencedNot yet evidenced: a research gap, not evidence of absenceVerified with a scope note: the Atlas bounds this cell to a named programme or route; open the dossier or record

Of 40 institution-by-hub cells, 27 are verified, 12 partial and 1 not yet evidenced.

All 175 institutions with evidence

Recently observed

The latest observations, by institution

The three institutions with the most recent observations in the register. Updates feed: RSS, JSON.

All release notes and observations

Deutsche Bank

39 observations in the register; latest

SIG-1427 Official source Deutsche Bank

Deutsche Bank lists a 15 December 2026 to 16 January 2027 early recruitment application window for 2027 Summer Internships in Fixed Income and Currencies and Investment Banking and Capital Markets across New York, Chicago and San Francisco.

United States / New York; Chicago; San Francisco · 2027 Summer Internship Early Recruitment · US early recruitment window · Observed 2 November 2026 · Official employer primary source directly reviewed 2026-11-02 · Source SRC-508 (careers.db.com) · source record

KPMG Singapore

2 observations in the register; latest

SIG-1424 Official source KPMG Singapore

KPMG Singapore's 2027 Organisation Transformation Graduate Associate vacancy lists a bachelor's degree in HR, Business, Psychology, Sociology, Communications, Economics, Engineering, Data Analytics, Social Sciences or a related discipline; leadership experience is described as an advantage.

Singapore · 2027 Consulting, Organisation Transformation Graduate Associate · Degree discipline and leadership signal · Observed 2 November 2026 · Official employer primary source directly reviewed 2026-11-02 · Source SRC-507 (careers.kpmg.com.sg) · source record

PJT Partners

10 observations in the register; latest

SIG-1422 Official source PJT Partners

PJT Partners says its Summer Analyst and Summer Associate programmes are generalist programmes and do not have a separate placement process within a business.

Global · Summer Analyst and Summer Associate · Generalist programme placement · Observed 2 November 2026 · Official employer primary source directly reviewed 2026-11-02 · Source SRC-505 (pjtpartners.com) · source record

New research publication · Observed 23 September 2026

AI & Elite Hiring Observatory

What do selective employers officially say about candidate AI use? A source-derived, stage-specific analysis of a purposive 67-employer check, with precise denominators and explicit limitations.

Read the AI & Elite Hiring Observatory · Examine the official-source tracker

Employer evidence profiles

AI rules, by employer and selection stage

Read source-bounded employer evidence, with official guidance, programme-specific limitations and dated observations. These are not complete application guides.

Read the AI & Elite Hiring Observatory