LinkedIn Trust & Safety

Job seekers cannot tell a real opportunity from a convincing scam

Recruitment fraud has grown into an everyday risk on professional networks. The people most exposed are the ones with the least leverage — candidates who need the role.

This is a self-initiated concept project. It was never deployed by LinkedIn.

• 40%

  rise in job fraud, 2021–2024

  Measured evidence

Small, direct, documented

Evidence came from scam artefacts collected in the wild and from talking to people who had encountered them.

• 2

  recorded scam interactions

  Measured evidence
• 11

  survey respondents

  Measured evidence
• 3

  personas synthesised

  Measured evidence
• 6

  features proposed

  Measured evidence

A sample of this size is directional, not statistically representative — the design decisions treat it as such.

Trust fails at the decision moment, not at the report moment

Reporting tools help after harm. Candidates needed the signal earlier — while reading the post, the company page and the recruiter profile.

Urgency is the scammer's main lever, so the counter-design has to be legible in seconds.

Layered verification instead of one badge

• Verified job seeker profile badge
• Experience approval interface
• Verified company page
• Job post with a visible trust score
• Scam alert notification
• Scam report form

Trust signals only work when they are consistent

The components share one verification grammar — the same badge, the same score treatment, the same warning hierarchy across profile, company and post surfaces, so the signal stays readable wherever it appears.

Assumptions, clearly labelled

The following figures are projections used to frame the business case for the concept. The solution was not deployed, so none of these are measured product performance.

• +250%

  verified profiles

  Projected
• +180%

  trust score adoption

  Projected
• +65%

  engagement

  Projected
• +120%

  premium conversions

  Projected
• −70%

  fraud reports

  Projected
• −85%

  fake postings

  Projected
• −60%

  complaints

  Projected
• −45%

  account suspensions

  Projected

Projected / expected impact — assumptions, not measured results.

Safety design is decision support

The valuable output was not the badge. It was deciding where in the journey a person is most vulnerable, and making the evidence available exactly there.