Autonomous AI Agents: Keeping Them Secure and Under Control

The UAE has embraced AI with enthusiasm, and it is now entering a thoughtful new phase. The biggest question for businesses may no longer be how to get AI working, but how to make sure it never works beyond its authority. With autonomous AI agents becoming part of daily operations, many organisations are asking how to keep them helpful, secure and under control. It is a healthy sign of maturity.

What Are Autonomous AI Agents?

Across the country, organisations are moving from experimenting with generative AI to deploying autonomous AI agents. These are systems that can retrieve information, interact with other systems, and perform tasks with limited human involvement. They promise faster work and fewer routine steps.

The Numbers Behind the Shift

Research from Proofpoint’s 2026 AI and Human Risk Landscape report found that 80% of UAE organisations are piloting or rolling out autonomous AI agents. The same research shows that 41% of UAE organisations have already experienced an AI-related security incident. A separate study found that 73% of UAE chief information security officers (CISOs) consider generative AI a security risk, up 19 percentage points from the previous year.

From Adoption to Accountability

Emile Abou Saleh, VP, Emerging Markets at Proofpoint, said every company is adopting AI, and that boards and CEOs are no longer asking whether to adopt it, but how to do so in a way that is safe, secure, fast and scalable.

A Simple Example

Emile Abou Saleh warned that autonomous AI agents could give users information or access to data they would not normally be allowed to see. Imagine an employee asking for a colleague’s salary. Through standard systems, access is usually limited to authorised staff. But an AI agent connected to company information might retrieve the same details if proper safeguards are missing.

The question, then, is not only whether the technology can find the information, but whether it should. Organisations need to understand why an agent was created, what responsibilities it has, what information it needs and where its authority must end.

People, Data and AI Agents

Traditional security focuses on employees and sensitive information. Autonomous AI adds a third element: the systems acting on employees’ behalf. Abou Saleh said businesses must consider people, data and AI agents together.

Existing access controls and data protection policies cannot be assumed to cover the risks agents create. The challenge grows because cybercriminals also use AI to do in less time work that once took hours, weeks or months.

Three Areas to Watch

Kenan Abu Ltaif, regional lead for the Middle East and Turkey at the company, said the UAE’s rapid adoption makes this especially pressing, noting that 41% of UAE organisations have experienced an AI-related security incident. He pointed to three areas:

  • Technology: tools that give visibility into how AI is being used
  • Processes: clear policies on when and why AI should be deployed
  • People: employees who understand their responsibility

As Kenan Abu Ltaif put it, every employee needs to understand their responsibility when using or creating an AI agent, and agents must be connected to the right data and set up for their intended purpose.

Looking at Intent

Abou Saleh also described a promising approach: looking beyond fixed rules to understand intent, and Abu Ltaif added that security must now protect both people and AI systems. Security systems assess not only what an agent is trying to do, but why. They can then decide whether to allow an activity, block it, or adjust it, based on its purpose and risks.

Signs of Progress Across Sectors

The move towards governance is visible elsewhere:

  • At ADIPEC 2026, AIQ CEO Dennis Jol spoke about coordinating agents, verifying their identities and governing their activity in the energy sector. AIQ presented Genesis, an AI-native operating system that brings AI models, applications and agents together in an industry-specific environment
  • On 9 November, Tarahum Charity Foundation launched Rahma, an AI agent that guides website visitors, donors and customers through its services

Why This Conversation Deserves Appreciation

Thoughtful attention to autonomous AI agents deserves appreciation for several reasons:

  • It treats AI safety as a business priority, not an afterthought
  • It encourages clear rules about what agents may and may not do
  • It protects people as well as data
  • It shows the UAE moving from ambition towards careful implementation

Conclusion

The UAE’s AI journey is entering a thoughtful chapter. By asking how much authority autonomous AI agents should have, and how to make sure they stay within it, businesses are building trust alongside innovation.

If your organisation wants guidance on setting clear, safe boundaries for AI, Horizontal8 is here to help. Reach out to Horizontal8 today, and let’s talk about your next step.

Frequently Asked Questions

How many UAE organisations are piloting or rolling out autonomous AI agents? 

According to Proofpoint’s research, 80%.

How many have had an AI-related security incident? 

41% of UAE organisations.

What is the main risk with AI agents? 

They could give users information or access they are not authorised to have if safeguards are missing.

What are the three areas to focus on? 

Technology, processes, and people.

What does looking at intent mean? 

Security systems assess why an agent is acting, not only what it is doing, and then decide to allow, block, or adjust the activity.

Free Brand Audit