AI-Written CVs and Candidate Fraud: How UAE Employers Can Verify Tech Talent

A software engineering or cybersecurity vacancy attracts dozens of applications. The CVs look polished, the keywords match the job description and the experience appears right. Then the interviews start, and some candidates struggle to explain the projects, tools and results listed on their own CVs.

For UAE employers competing for specialist technology professionals, that creates a real problem: how do you tell genuine expertise from a convincing document?

AI has made it easy to rewrite a CV, tailor it to an advert and rehearse interview answers. That helps qualified people communicate better, but it can also help others exaggerate or misrepresent their experience. The answer is not to reject AI-written CVs. It is to verify skills, experience and identity more carefully.

This guide covers what the research says, where traditional screening falls short and how to build a more reliable technology hiring process.

Why a Polished CV No Longer Proves Technical Ability

When many candidates use the same tools, presentation stops being a useful signal. It helps to separate three different situations:

• AI-assisted applications. A candidate uses AI to fix grammar, structure a CV or describe real achievements more clearly. This is not dishonesty.
• Exaggerated experience. A candidate claims expertise in tools they have barely used, or presents a team's work as their own.
• Fabricated identity or credentials. An applicant misrepresents who they are, where they worked or what they studied. In serious cases, someone else attends the interview on their behalf.

Each needs a different response. Well-written language alone should never decide an outcome, but material inconsistencies should trigger verification. The question to keep asking is whether the candidate can demonstrate the skills the job requires.

What the Research Says

Gartner's 2025 candidate surveys give useful context:

• 6% of 3,000 surveyed candidates admitted to interview fraud, meaning they posed as someone else or had someone pose as them.
• Gartner predicts that by 2028, one in four candidate profiles worldwide will be fake.
• In a separate survey, 39% of candidates said they used AI during the application process.
• Only 26% of candidates trust AI to evaluate them fairly.

Read these carefully. The 2028 number is a forecast, not today's rate. The 6% figure covers admitted interview fraud, not every CV exaggeration. None of the findings measures fraud in the UAE specifically.

The regional hiring market adds pressure. The Hays GCC Salary Guide 2026 reports that 90% of organisations had skills gaps in 2025, with demand highest for technical specialists and flexible staffing solutions. These are GCC-wide findings, not UAE-only, but they show why employers feel pressure to move fast on technical hires, which is exactly when screening shortcuts creep in. For a UAE-specific view of where demand sits, see our guides to top IT skills employers want in Dubai and jobs in demand in the UAE.

The practical takeaway: a CV is a useful starting point, but it is not proof of ability.

Seven Ways to Verify Technical Candidates

1. Define the role around outcomes, not a tool list

Start with what the successful hire must deliver. A cloud engineer, for example, might need to design secure infrastructure, automate deployments, monitor performance and handle incidents. That is a far better basis for assessment than a wish list of every tool you use.

Split requirements into essential skills the person must demonstrate and desirable skills they can learn after joining. This discourages keyword stuffing and helps you spot capable candidates whose previous stack differs slightly. It matters most for specialist work in software engineering and cloud transformation.

2. Ask screening questions that reveal real experience

Before progressing anyone, ask a few questions tied directly to the role.

For a software engineer:

• Which parts of the application did you personally design or build?
• What technical problem did you hit, and how did you solve it?
• Which languages and frameworks did you use in production?
• What would you change if you rebuilt it today?

For cybersecurity roles, ask about incident response, vulnerability management, access controls or a specific improvement they delivered. For ERP and CRM roles, ask which modules they configured, which integrations they supported and which business processes improved.

The aim is not to catch people out. It is to confirm that their experience is specific, relevant and consistent. This is also the moment to check availability, notice period and work-authorisation status.

3. Use a short, role-relevant assessment

A practical task is stronger evidence than a CV. Keep it proportionate to the seniority of the role. Examples:

• Software engineering: review a code sample, find a bug or explain a design decision
• Cloud: diagnose a sample infrastructure problem or propose a secure deployment
• Cybersecurity: analyse a fictional incident and prioritise the response
• Data analytics: interpret a small dataset and explain the conclusions, a skill we cover in our business intelligence and analytics hiring
• ERP: explain how a sample requirement would be met in the platform

Avoid long unpaid assignments that amount to free work. Because AI can also complete take-home tasks, follow up with a short live walkthrough where the candidate explains their reasoning, discusses alternatives or solves a related problem. Test job-relevant skills, not speed or familiarity with a testing format.

4. Build interviews around project ownership

Ask candidates to walk through a project from their CV in real depth:

• What problem was it solving?
• What was your individual contribution?
• Which technical decisions did you make, and why?
• What trade-offs or limits did you face?
• How did you measure the result?
• What would you improve?

If someone says they led a cloud migration, ask how they handled downtime, access controls, data integrity and rollback. People with real experience can usually go deeper on decisions and responsibilities. But nervousness or an unfamiliar interview format is not evidence of fraud, so judge technical reasoning, not polish.

5. Verify identity, employment history and qualifications

Verification should match the seniority and sensitivity of the role, and happen at the right stage, typically for shortlisted or final candidates. It may involve:

• Confirming identity matches the information provided
• Verifying previous employment through references or authorised channels
• Checking educational qualifications and professional certifications
• Resolving material inconsistencies across documents and answers

Some employers confirm only dates and job titles. Where detailed references aren't available, use other evidence such as work samples, verified credentials or a structured assessment. For remote interviews, a live identity check can be part of the process, provided you handle personal data lawfully. If a specific permit or employment process requires attested documents, confirm the requirement with the relevant authority rather than assuming it applies to every role.

Give candidates a chance to explain discrepancies before you decide. Date and title differences often have innocent explanations.

6. Cross-check the full journey, and know the warning signs

Compare the CV, application form, assessment, interview answers and references. A candidate who calls themselves the lead architect but can't explain the design or deployment deserves further questions, though that alone doesn't prove misrepresentation.

These are prompts for a follow-up, not proof of fraud:

• Dates or titles that differ between the CV and a reference
• Difficulty explaining work the candidate claims to have done alone
• Rehearsed or scripted-sounding answers that don't adapt to follow-up questions
• Reluctance to take part in a live identity or technical check
• Near-identical wording across several applicants

Focus on material inconsistencies, not writing style. Keep brief records of assessment results and verification decisions.

7. Protect your systems during onboarding

A fraudulent hire in a role with access to source code, cloud infrastructure or customer data creates risks far beyond a bad hiring decision. For sensitive technology roles:

• Complete identity and employment checks before granting sensitive access
• Use approved accounts and multifactor authentication
• Apply least-privilege access, so people get only the permissions they need
• Grant elevated permissions only when justified and authorised
• Review access as responsibilities change, and remove it promptly when someone leaves

These controls belong in your wider security and onboarding process. They shouldn't become excessive surveillance. If you need support with the security side of hiring, see our IT infrastructure and cyber security specialisation.

A Simple Technical Scorecard

A scorecard helps hiring managers compare candidates on evidence instead of impressions. Treat the weights as a starting point and adapt them to each role.

Criteria and weights

• Technical knowledge: 25%
• Practical problem-solving: 25%
• Project ownership: 20%
• Communication and reasoning: 15%
• Relevant experience: 15%

Score each criterion from 1 to 5

• 1: cannot demonstrate the capability, even with clarification
• 2: limited understanding, needs substantial guidance
• 3: meets the requirement with a workable explanation or solution
• 4: strong independent judgement and clear trade-offs
• 5: exceptional depth, sound judgement and clear rationale

To get a weighted score, multiply each score by its weight and divide by five, then add the results to reach a total out of 100. A candidate scoring 4 out of 5 on technical knowledge earns 20 of the available 25 points.

Record a short piece of evidence for every rating. Define pass-or-fail essentials before interviews begin, because a high total shouldn't make up for missing a genuinely mandatory requirement. The scorecard supports professional judgement and does not replace it.

What to Avoid When Screening AI-Assisted Applications

More checks are not always better. A badly designed process excludes good candidates, slows hiring and damages trust.

• Don't reject candidates just for using AI. Evaluate whether their claims are accurate and their skills are real.
• Don't treat AI-text detectors as proof. They produce false positives and false negatives, so a detector score should never be your only evidence.
• Don't stretch the process. Extra interviews, repeated tests and long unpaid tasks push strong candidates away. Use the fewest stages that give a reliable decision.
• Do set clear AI rules. Tell candidates whether AI is allowed in applications, coding exercises and take-home tasks, and state up front when independent work is required. Gartner also advises employers to explain how they define acceptable AI use and to be open about their fraud detection efforts.
• Don't over-collect personal data. Identity and background checks should be relevant, proportionate and compliant with data protection law. The UAE federal framework includes Federal Decree-Law No. 45 of 2021, and certain free-zone regimes have their own rules. Explain what you collect and why, restrict access, follow retention and deletion practices, and take qualified advice where needed.
• Don't forget employment compliance. Verification doesn't replace work-authorisation, permit or onboarding requirements. Recruitment fees must not be charged to prospective employees where the law prohibits it. Our guide to UAE Labour Law explains the basics, and you should confirm current obligations with the relevant authority.

Source Candidates Beyond Job Adverts

Better screening is only half the answer. You can also improve the quality of your applicant pool:

• Employee referrals: your team knows capable people who aren't applying
• Direct sourcing: approach professionals whose experience matches the vacancy
• Talent pipelines: build relationships before roles become urgent
• Specialist recruitment partners: firms with deep technical knowledge can pre-screen against your requirements

A specialist IT recruitment partner in Dubai can help with hard-to-fill roles, and for project-based needs, contract staffing can give you specialist skills without a permanent hire. Our overview of the benefits of outsourcing in the UAE explains when that model makes sense, and our guide to recruitment agency fee structures in Dubai will help you compare costs. If you'd like to talk through your own vacancies, you can contact our team or see how we work on our about page.

Outsourcing doesn't remove fraud risk, because agencies receive inaccurate and AI-assisted applications too. Ask any provider:

• How do you assess technical experience before submitting a candidate?
• Which identity, employment and qualification checks do you perform?
• How do you handle duplicate applications or material inconsistencies?
• Which verification responsibilities stay with us?
• What are the fees, replacement conditions and guarantee terms?

Agreeing these answers upfront prevents gaps between the agency's checks and yours.

Quick Checklist for UAE Technology Employers

• Define essential skills and expected job outcomes
• Separate mandatory from desirable requirements
• Ask specific screening questions about real projects
• Use a short, role-relevant assessment with a live follow-up
• Run structured interviews with consistent scoring
• Record the evidence behind each assessment
• Verify identity, employment history and qualifications where appropriate
• Resolve material inconsistencies fairly before deciding
• Apply access controls during onboarding
• Protect candidate data and follow privacy requirements
• Communicate acceptable AI use before assessments
• Use referrals, direct sourcing or a specialist partner alongside job adverts

Frequently Asked Questions

Are AI-written CVs a problem for UAE employers?

Not by themselves. Candidates can use AI to improve grammar and structure without misrepresenting their experience. The concern is materially exaggerated or fabricated claims, so assess real capability instead of rejecting applications for how they were written.

Can employers reliably detect an AI-written CV?

Not with certainty. Detection tools make mistakes in both directions. A more dependable approach is to verify experience through structured interviews, practical assessments, references and supporting documents.

How common are fake candidates?

Gartner found that 6% of surveyed candidates admitted to interview fraud and predicts one in four candidate profiles worldwide will be fake by 2028. The second figure is a forecast, and neither is a measured UAE rate.

How can we verify candidates interviewed remotely?

Use proportionate live identity checks, verify employment history and qualifications, contact references through suitable channels and ask candidates to explain their own project work. Tell candidates what you're checking and why. An in-person final interview can help for some roles but isn't required for all.

What is the best way to assess technical skills?

Combine methods: a short work sample, a structured technical interview and a review of relevant experience. No single test is enough. Keep assessments job-relevant, apply the same criteria to every candidate and give people a fair chance to show their ability.

Should we stop keyword screening?

No. Keyword screening helps organise applications, but it shouldn't be the only basis for deciding who is qualified. Pair it with specific screening questions, practical evidence and structured interviews.

Can a recruitment agency prevent fraudulent applications?

A good specialist agency can source candidates and run agreed checks, but it can't guarantee every application is accurate. Ask what screening it performs, clarify which checks stay with you and keep oversight of the final hiring decision.

Final Thoughts

AI has changed how people prepare applications, but it hasn't removed the need for real expertise. A polished CV helps a candidate communicate. It can't replace the ability to solve problems, explain decisions and show relevant skills.

The best response is to define role-specific requirements, assess technical skills early, state your AI rules, interview in a structured way, verify shortlisted candidates fairly and protect sensitive systems during onboarding. The goal isn't to treat every AI-assisted application as suspicious. It is to build a process where qualified professionals can prove their ability and hiring decisions rest on credible evidence.

Hiring software engineers, cybersecurity professionals, cloud specialists or other technology talent in the UAE? Get in touch with Staff Connect to discuss your requirements and the screening approach that suits your roles.