AI adoption in the UAE is moving quickly from ambition into practical use. Recent agreements involving OpenAI, G42, ADNOC and Abu Dhabi Government entities show how AI is being applied across government, energy, investment, enterprise and education.
At the same time, the UAE is responding to significant cyber security pressure. The UAE Cyber Security Council has reported hundreds of thousands of attempted cyberattacks, including threats supported by AI. These figures refer to detected attempts, not successful breaches.
The two developments are not evidence that greater AI adoption is causing more cyberattacks. But they do highlight an important challenge for organisations adopting AI: innovation and cyber resilience need to progress together
AI is moving into everyday operations in the UAE
In November 2026, four agreements were signed between OpenAI and Abu Dhabi-based organisations, covering government, energy and investment, alongside wider AI adoption across government, enterprise and education.
One agreement brings together ADNOC, G42 through its subsidiary Inception42, and OpenAI. The collaboration is intended to integrate AI into ADNOC operations, supporting areas including operational, investment and commercial decision-making. OpenAI models will use in-country inference infrastructure in the UAE.
The Abu Dhabi Government and OpenAI have also launched a Machine-Readable Government initiative. This takes defined government policies and regulations and converts them into formats that AI systems can process.
Its first pilot is already in production with the Abu Dhabi Agriculture and Food Safety Authority, where AI supports regulatory compliance assessments relating to food imports. Importantly, government officials remain responsible for oversight and final decisions.
These developments are at different stages. Some are operational, while others represent longer-term implementation commitments. But collectively they show something important: AI is moving closer to the data, systems and processes that organisations rely on every day.
As that happens, the secure integration of AI with organisational systems and data becomes increasingly important.
Cyber threats are evolving alongside AI
The UAE’s AI investment is developing against an active cyber threat landscape. Reporting based on information from the UAE Cyber Security Council has put the level of cyberattack attempts at around 800,000 per day during a period of heightened activity. AI has reportedly been used to support phishing, identify vulnerabilities and accelerate the development of malicious code.
The Cyber Security Council has also reported coordinated attacks targeting sectors including aviation, energy and education. National cyber systems detected and contained these attacks before they could disrupt critical systems and services.
The scale and speed of this activity reflect a wider change in the threat landscape. AI is helping attackers create more convincing campaigns and exploit vulnerabilities faster, giving organisations less time to detect and respond. As AI becomes more integrated into everyday operations, the security measures protecting data, identities and systems need to develop alongside it.
What does this mean for organisations adopting AI?
AI does not become a cyber security problem simply because an organisation starts using it.
The challenge comes when AI is introduced without understanding what it can access, which systems it connects to and what actions it is permitted to take.
For example, an AI tool connected to organisational systems could potentially have access to business data, email, documents or customer information. An AI agent may also be able to take actions across those systems.
That makes identity, permissions and data governance important parts of AI adoption.
Some of the areas organisations should consider include:
- Access to data: AI applications should only have access to the information required for their intended purpose.
- Unapproved AI tools: Employees need clear guidance on which platforms they can use and what information can be shared with them.
- AI integrations: Connections to email, documents, customer platforms and other business systems should be reviewed before deployment.
- Permissions: Users, applications and AI agents should follow least-privilege principles.
- Visibility: Organisations need to understand how AI tools access and use business information.
- Human oversight: Sensitive or consequential actions should have appropriate human approval.
The aim is not to restrict AI use. It is to make sure organisations remain in control as their use of AI develops.
Building security into AI adoption
Secure AI adoption starts before a new tool or agent is deployed.
Set clear rules for AI use
Organisations need practical policies covering approved tools, appropriate use and the types of information that should not be entered into public AI platforms.
The goal should be responsible adoption, not simply restriction.
Control identity and access
Multi-factor authentication, role-based access control and least-privilege permissions can help ensure people, applications and AI agents only access what they need.
Permissions should also be reviewed as roles, integrations and AI use cases change.
Understand and govern your data
Before giving AI access to organisational information, businesses need to understand where that data sits, how sensitive it is and who should be able to access it.
Clear data classification and governance provide stronger foundations for AI adoption.
Review AI integrations
Connecting an AI application to another business system should involve more than checking whether the technology works.
Organisations should understand what information an integration can read, what actions it can perform and whether those permissions are appropriate for the intended use case.
Monitor what is happening
Visibility remains essential as AI becomes more integrated with everyday technology.
An integrated Microsoft Security Services approach can help organisations protect identities, devices, data, email, applications and cloud environments while maintaining visibility across the wider technology estate.
Keep people in control
The approach being taken by the Abu Dhabi Government provides a useful example. AI can support regulatory assessments, but officials retain oversight and accountability for the final decision.
The same principle should apply when businesses introduce AI into consequential processes. AI can augment people and support faster decision-making, while accountability remains with the organisation.
How Zenzero can help
Secure AI adoption is not about adding security after implementation. It starts with the environment AI is being introduced into.
Zenzero brings together expertise across Microsoft technologies, cyber security, data and managed IT to help organisations build the foundations needed for responsible AI adoption.
This can include assessing AI and data readiness, reviewing identity and access controls, strengthening the wider cyber security environment, understanding how AI connects with organisational data and considering appropriate monitoring and human oversight.
The objective is practical: help organisations unlock the value of AI while maintaining control over their data, systems and security.
Final thoughts
The UAE provides a clear example of how quickly AI is moving into practical use across government services, energy, education and enterprise. Its evolving cyber security landscape is an equally important reminder that innovation and resilience need to progress together.
As AI becomes more closely connected to organisational data, systems and decisions, security cannot be added afterwards. Identity and access controls, data governance, monitoring and human oversight need to form part of adoption from the beginning.
AI creates significant opportunities for organisations. Building it on secure, governed foundations is how those opportunities can translate into lasting business value.
If your organisation is considering how to adopt AI securely, contact Zenzero to discuss the right next step.
