Using AI in hiring lawfully in the United Arab Emirates
Reviewed by Mellow Editorial Team, HR & payroll content team
AI hiring tools can speed up screening and reduce administrative load, but UAE employers must apply them within a clear legal and ethical framework. This article explains what that framework looks like today and where the practical risks sit.
What the UAE legal landscape actually covers
The UAE does not yet have a single statute that regulates AI in recruitment specifically. Instead, several overlapping frameworks apply.
Federal Decree-Law No. 33/2021 (the Labour Law) prohibits discrimination in hiring on grounds including gender, race, religion, national origin and disability. If an AI tool produces outcomes that systematically disadvantage a protected group — even unintentionally — the employer is exposed to a discrimination claim regardless of whether a human made the final call.
The UAE Personal Data Protection Law (Federal Decree-Law No. 45/2021, in force with its executive regulations) applies whenever you collect, process or store personal data. A CV, a video interview recording, a psychometric score — all of it is personal data. Lawful processing requires a valid legal basis, a clear purpose and appropriate retention limits. Candidates have rights to access and, in some circumstances, to object to processing.
The UAE's AI Ethics Guidelines, published by the Office of AI, are not legally binding but signal the direction of regulatory thinking. They emphasise transparency, accountability and avoiding bias. Regulators, courts and labour authorities are likely to reference them when evaluating employer conduct.
Where bias enters AI hiring tools
Bias is the most significant practical risk. Most commercial AI screening tools were trained on historical hiring data. If that data reflects past patterns — certain universities, certain demographic profiles, certain writing styles associated with successful hires — the model will reproduce those patterns at scale.
Common problem areas:
- Resume screening models that down-rank non-linear career paths or names associated with particular ethnic or national backgrounds.
- Video interview analysis tools that score candidates on vocal tone or facial expression, which can disadvantage non-native speakers or people with certain disabilities.
- Automated ranking systems that use proxies (such as postcode or institution type) that correlate with protected characteristics.
Audit your tools before deploying them. Ask vendors directly whether they have conducted bias testing, on what demographic data, and whether they will share the results. A vendor unwilling to answer those questions clearly is a vendor worth reconsidering.
Transparency obligations toward candidates
Even where the law does not yet mandate specific disclosures about AI use, transparency is both good practice and consistent with the spirit of the Personal Data Protection Law.
At a minimum, candidates should know:
- that automated tools are used in the screening or assessment process
- what type of data is collected (CV text, video, psychometric responses)
- how long data is retained and whether it is shared with third parties
- that a human has final authority over hiring decisions
This last point matters legally. Where a decision has significant consequences for a person — and rejection from a job clearly qualifies — there is a strong argument under data protection principles that a human should review and be accountable for the outcome, rather than leaving it entirely to an automated decision.
Update your job application privacy notice to reflect AI use. Many UAE employers are still running notices written before AI screening tools existed.
Human oversight in practice
AI tools should inform decisions, not make them. This is not just a compliance position — it also produces better outcomes.
A workable model for most employers:
1. Use AI to filter for minimum objective criteria (specific qualifications, required experience, language requirements) where those criteria are genuinely job-relevant and applied consistently.
2. Have a recruiter or hiring manager review any candidate the AI would reject at the first stage, particularly where the candidate is close to the threshold.
3. Document the criteria the AI applies and why they are relevant to the role.
4. Retain records of decisions — both the AI output and the human review — so you can respond to a challenge or a data access request.
Documentation is your protection. If a rejected candidate claims discrimination, your ability to demonstrate that the criteria were objective, consistently applied and reviewed by a human will determine how defensible your position is.
Due diligence on third-party vendors
Most UAE employers use off-the-shelf AI hiring products built outside the UAE. That creates additional obligations. Under the Personal Data Protection Law, transferring personal data outside the UAE to a vendor requires either that the destination country meets an adequacy standard or that appropriate contractual protections are in place.
Check your vendor contracts for:
- data processing agreement terms that assign responsibility clearly
- confirmation of where data is stored and processed
- breach notification commitments
- their own compliance with applicable data protection laws
Vendor due diligence is not a one-off exercise. AI tools are updated regularly, and a model that passed your initial review may behave differently after a vendor update. Build periodic review into your procurement and HR governance processes.
---
Run HR and payroll in UAE with Mellow
Mellow brings HR, payroll and 12 AI agents into one platform — built to handle UAE properly, with payroll included, from £4 per employee per month. The AI agents don't just answer questions; they generate contracts, run cost estimates and draft letters for you.
[Start a free trial →](/register)