Cisco highlights growing data security risks as Shadow AI adoption expands

Cisco has highlighted the growing cybersecurity risks linked to shadow AI, as businesses increasingly use artificial intelligence tools outside approved enterprise systems.

As organisations across the UAE expand their use of AI, employees are turning to generative AI applications to improve productivity and automate daily tasks. However, using these tools without proper security controls can expose sensitive business information and create security gaps that are difficult to detect and manage.

According to research cited by Cisco, organisations with high levels of shadow AI experienced data breach costs averaging $670,000 more than those with little or no shadow AI. The research also found that 97% of organisations that experienced an AI-related security incident lacked proper AI access controls.

Fady Younes, Managing Director, Cybersecurity, Cisco Middle East, Türkiye, Africa, Caucasus and Central Asia (METAC), said: “AI is helping businesses work more efficiently and unlock new opportunities. However, without the right security measures, unauthorised AI tools can expose sensitive data and create risks that organisations may not be aware of. Businesses need clear policies, strong access controls and better visibility into how AI is used. This will help them protect their systems while allowing employees to use AI safely and effectively.”

Strengthening security as AI adoption expands

As AI tools become more integrated into daily business operations, Cisco emphasises the importance of extending existing cybersecurity practices to address emerging risks. This is particularly important as AI systems gain greater access to enterprise data and become capable of performing tasks independently.

Cisco outlines three key priorities to help organisations manage shadow AI risks:

• Improve visibility into AI usage: Identify and monitor AI applications used across the organisation, including unauthorised tools, personal accounts and browser extensions. This can help security teams detect potential risks and understand how employees are using AI.
• Strengthen access controls and data protection: Limit access to sensitive information and apply clear data protection policies. Organisations should ensure that AI tools and systems only access the information needed to perform their intended tasks.
• Establish clear AI governance policies: Introduce practical guidelines for AI use, supported by regular monitoring and risk assessments. Providing employees with approved AI tools can also reduce reliance on unauthorised applications.

Cisco highlights that securing AI adoption does not require organisations to restrict innovation. Instead, businesses should provide employees with trusted tools and clear guidance to support productivity without compromising security.

While AI introduces new cybersecurity challenges, it also offers opportunities to strengthen security operations. Cisco emphasises that organisations must balance innovation with effective safeguards, ensuring they can benefit from AI while maintaining visibility and control.

By integrating AI governance into their wider cybersecurity strategies, organisations can better protect sensitive information, manage emerging threats and support responsible AI adoption.