Back to blog
Digital HR & AI

Using AI in HR Without Losing Trust

Where AI is safe to use in HR, where it needs controls and where to stop, plus the data rules in Saudi Arabia and Egypt, vendor questions and a 30-day plan.

By Mohamed ShaheenNovember 18, 20268 min read

AI can save an HR team hours every week. It can draft a job ad in a minute, sort hundreds of applications and summarise a survey with thousands of comments. But HR works with people's jobs, pay and careers. If people feel a machine is judging them in secret, you lose something that is much harder to rebuild than a process: their trust.

This guide is practical. It shows where AI is safe to use in HR, where it needs controls and where to stop. It also covers the rules that apply in Saudi Arabia and Egypt, and a 30-day plan to put it all in place.

The trust gap in numbers

A 2025 Gartner survey of almost 3,000 job candidates found that only 26% trust AI to evaluate them fairly, even though 52% believe AI is already screening their applications. A quarter said they trust an employer less when they know it uses AI to evaluate them.

At the same time, many HR teams are only now starting. SHRM's State of AI in HR 2026 found that 62% of organisations use AI somewhere in the business, but only 39% use it in HR. If you are early, that is good news: you can set the rules before bad habits form.

The rules that apply to you

Saudi Arabia

  • The Personal Data Protection Law (PDPL) covers the candidate and employee data that AI tools process. Its Implementing Regulations require explicit consent when a decision is based entirely on automated processing.
  • SDAIA published draft amendments to those regulations in October 2026. They would narrow this rule to decisions based solely on automated processing that affect people's rights or interests. Until they are final, follow the current text.
  • SDAIA's AI Ethics Principles set seven principles, including fairness, transparency and explainability, and accountability.
  • A draft Responsible AI Policy was put to public consultation in April 2026. It would add registration with SDAIA and stronger oversight duties, starting with large companies and high-risk sectors. Watch for the final version.

Egypt

  • The Personal Data Protection Law (Law No. 151 of 2020) is now fully in force. Its Executive Regulations were issued on 1 November 2025, and the one-year grace period that followed has now ended.
  • Sending personal data outside Egypt needs a licence or permit from the Personal Data Protection Centre. This matters because many AI tools store data on servers abroad.
  • The Egyptian Charter for Responsible AI (2023) is not binding, but it sets clear expectations: human-centred design, transparency, fairness, accountability and security.

If you hire in Europe

  • Since February 2025, the EU AI Act bans AI that infers emotions at work, including during recruitment.
  • AI used for recruitment, promotion, termination, task allocation or performance monitoring is classed as high-risk. After the 2026 "Digital Omnibus" changes, those obligations apply from 2 December 2027.

A traffic-light map for AI in HR

LightTypical usesRule
Green: goDrafting job ads and interview questions, first drafts of policies and letters, scheduling, an FAQ assistant built on your approved policiesA person checks the output before it is used
Amber: controls firstCV screening and ranking, skills matching, attrition-risk flags, interview notes and summaries, help with writing performance reviewsTell people, keep a named human decision-maker and check results for fairness
Red: stopAI making hire, fire, pay or promotion decisions on its own; emotion or facial analysis in interviews; scoring people from keystrokes or webcams; pasting personal data into public chatbotsDo not use, whatever the vendor promises

Five rules that protect trust

  1. Tell people. Say in plain language where AI is used: in the job ad, the offer process and the employee handbook. Secret use is the fastest way to lose trust.
  2. A human owns every decision. Name the person who decides, and make sure they actually read the profile instead of approving the AI ranking.
  3. Give people a way to challenge. Candidates and employees should be able to ask for a human review and get an answer within a set time.
  4. Protect the data. Use company-approved tools only, keep personal data out of public chatbots, and know where your vendor stores the data.
  5. Check for fairness every quarter. Compare who passes each stage by gender, nationality, age group and university, and read a sample of rejected applications yourself.

Sample wording for your job ads

We use AI tools to help us organise applications and match skills to the role. A member of our team reviews every application before any decision is made. If you would like a person to review your application again, email careers@yourcompany.com.

Short, honest and easy to find. Adapt the address and add a similar note to your employee handbook for any AI used with current staff.

Questions to ask any AI HR vendor

  • Where is our data stored, and can it stay in Saudi Arabia or Egypt if the law requires it?
  • Is our data used to train your models? Can we opt out?
  • What exactly does the tool score, and can it explain each result?
  • How do you test for bias, and can we see the latest results?
  • Can our team override any result, and is the reason logged?
  • What happens to our data when the contract ends?

A 30-day plan to get started

  • Week 1: list every AI tool your HR team and managers use, including personal accounts on public chatbots.
  • Week 2: sort each use into green, amber or red. Stop anything red straight away.
  • Week 3: write a one-page AI use note for HR and add the candidate and employee wording.
  • Week 4: train HR and hiring managers on the five rules, and book the first quarterly fairness check.

The AI use note can sit alongside your core policies. If you are still building those, start with Your First 10 HR Policies. For the values behind all of this, read Ethical AI and the Future of People Decisions in HR.

Frequently asked questions

Do we need consent to screen CVs with AI?

In Saudi Arabia, the current PDPL Implementing Regulations require explicit consent when a decision is based entirely on automated processing. The simplest safe approach in both countries is to keep a human in every decision, tell candidates how AI is used and get legal advice on your exact setup.

Can our HR team use ChatGPT or similar tools?

Yes, for drafting and research, through accounts the company approves. Do not paste names, salaries, ID numbers or other personal data into public tools.

Will AI replace HR jobs?

It will change them. Routine admin will shrink, while judgement, coaching and employee relations matter more. Reskilling the Workforce for the AI Era covers how to prepare.

Trust is the real productivity tool

AI makes HR faster. Trust makes HR work. Be open about where you use AI, keep a person responsible for every decision and check the results regularly. Then people will see AI as help, not as a hidden judge.

If you would like help choosing tools, writing your AI use note or reviewing a vendor, you can work with me directly.

Note: This article summarises the rules as of November 2026 and is not legal advice. AI and data protection rules are changing fast in both countries; check the current requirements before you act.

Last updated: 18 November 2026

References

Share

Keep reading