Do You Have to Tell Customers They're Talking to AI? Law and Ethics in Saudi Arabia and the Gulf
Must you tell customers they are chatting with AI? What SDAIA, the UAE AI Charter and the EU AI Act say, plus ready disclosure lines and handover rules.
In Saudi Arabia, SDAIA's AI Ethics Principles and Generative AI Guidelines call for transparency and telling people when AI is being used. In the UAE, the AI Charter stresses transparency and human oversight, and in Europe the AI Act has required since August 2, 2026 that people be told they are interacting with an AI system. The safest practice: disclose in one sentence in your welcome message and allow a handover to a human at any time.
01The short answer: what do the laws and principles say?
| Reference | What it says in brief | Nature |
|---|---|---|
| AI Ethics Principles, SDAIA (September 2023) | Seven principles, including transparency and explainability, privacy and accountability | Principles and guidelines |
| Generative AI Guidelines, SDAIA (2024) | Clearly inform the public when generative AI capabilities are used and when content is machine-generated | Guidelines |
| Implementing Regulations of the Personal Data Protection Law (Saudi Arabia) | Explicit consent for decisions based entirely on automated processing, and an impact assessment for that processing | Binding law |
| Charter for the Development and Use of AI, UAE (June 2024) | Twelve principles, including transparency, human oversight and data privacy | National policy |
| EU AI Act, Article 50 | Inform people that they are interacting with an AI system unless this is obvious | Binding since August 2, 2026 for anyone serving users in the EU |
This is a general summary, not legal advice; check the official text or consult a lawyer. Sources: SDAIA, the UAE official legislation portal, and texts and analyses of the EU AI Act, September 2026.
02Saudi Arabia: clear principles and a binding data law
As of September 2026, chatbot disclosure in Saudi Arabia comes mainly from principles and guidelines: SDAIA's transparency and explainability principle, and the Generative AI Guidelines, which ask that the public be informed when it is used. These carry no direct penalties, but they are the benchmark against which responsible practice is measured.
- Automated reply to a question about price or an appointment: disclosure is recommended practice.
- A decision that affects the customer and is made fully automatically, such as rejecting an installment or refund request: here the Implementing Regulations of the Personal Data Protection Law apply, requiring explicit consent and an impact assessment, and the safest option is to have a staff member review it.
- Collecting data through the agent: tell the customer the purpose and respect their consent, as with any other channel.
03The UAE and the rest of the Gulf
The UAE Charter for the Development and Use of Artificial Intelligence, issued in June 2024, includes transparency and human oversight among its twelve principles and stresses that humans must keep judgment and control over AI systems. In the other Gulf states, as far as we have found, the clearest framework is personal data protection and consumer protection law, not a text specific to chatbots.
If you operate in more than one country, apply the highest standard everywhere: disclosure in the welcome message and a handover to a human that is always available. It is easier than managing a separate policy per country.
04If you have customers in Europe
The EU AI Act applies the transparency obligations in Article 50 from August 2, 2026 to anyone who provides or deploys AI systems for users in the European Union, even if the company is based outside Europe. Systems that interact with people must be designed so that people know they are dealing with AI, unless this is obvious from the context.
- It concerns you if: you run a hotel, travel company or store that serves customers living in EU countries.
- Fines: up to EUR 15 million or 3% of global annual revenue, whichever is higher.
- Recent amendments (the Digital Omnibus package) extended some deadlines for labeling generated content, but did not postpone the duty to inform users in a conversation, according to legal analyses published in July 2026.
Source: McCann FitzGerald analysis, "One Month to Go: EU AI Act Transparency Compliance", July 2026.
05Why disclosure pays off even when it isn't mandatory
- Trust: a customer who later finds out they were talking to a bot feels deceived, even if the service was excellent.
- Accurate expectations: customers write more clearly and ask for a human when they need one.
- Protecting your reputation: a screenshot of a bot claiming to be "Ahmed from customer service" can go viral against you.
- Future-proofing: regulation is moving toward transparency, and those who start early won't have to change anything later.
Don't give the agent a human name and a staff photo and have it deny being automated when asked. A friendly name like "[Business name]'s smart assistant" is enough.
06Ready-to-use disclosure lines
One sentence in the first message is enough, without long warnings that put the customer off:
07When should the agent hand over to a human immediately?
- When the customer explicitly asks for a human, even in a single word.
- Anger, a repeated complaint or a threat to escalate.
- A financial decision concerning the customer: refund, exception, special discount, installments.
- A health or legal matter, or anything involving sensitive data.
- The agent fails to understand twice in a row.
Key takeaways
- SDAIA's principles and guidelines call for transparency and telling people when AI is used.
- Fully automated decisions that affect the customer fall under the data protection law and require explicit consent.
- The EU AI Act has required disclosure since August 2, 2026 for anyone serving users in Europe.
- A disclosure line in the welcome message and an always-available handover to a human is the safest practice.
Frequently asked questions
Is there a Saudi law that penalizes not disclosing a chatbot?
As of September 2026, disclosure appears mainly in SDAIA's principles and guidelines, while fully automated decisions and personal data are governed by the binding data protection law. Consult a lawyer about your case.
Does disclosure reduce sales?
Customers care about fast, accurate replies. Disclosing in a friendly sentence with a human option builds trust, while deception does more damage once it is exposed.
Should I give the bot a human name?
It is better to use a name that makes clear it is a smart assistant, such as "[Business name] Assistant". And if a customer asks whether it is a bot, it must answer honestly.
Does the EU law apply to my store in Riyadh?
It may, if you provide an AI system to users in the European Union. If all your customers are in the Gulf, the impact is limited, but disclosure is good practice either way.
- #SDAIA AI ethics principles
- #AI transparency
- #EU AI Act
- #chatbot welcome message