AI use in enterprises soars but brings surge in cyber risks

A surge of over 3,000% in enterprise use of artificial intelligence and machine learning tools has been revealed by Zscaler's ThreatLabz 2025 AI Security Report, with significant security concerns arising alongside this sharp growth.

The ThreatLabz 2025 AI Security Report analysed more than 536 billion AI transactions processed between February and December 2024 within the Zscaler Zero Trust Exchange platform. This study highlights real-world threats including AI-enhanced phishing, fraudulent AI platforms, and increased risks related to agentic AI and open-source models such as DeepSeek.

The report found that ChatGPT dominated usage, accounting for 45.2% of all AI/ML transactions, making it both the most popular and the most-blocked AI application. Grammarly and Microsoft Copilot followed as the second and third most-blocked tools, reflecting widespread enterprise concerns about data leakage and unsanctioned use of these platforms.

"We had no visibility into [ChatGPT]. Zscaler was our key solution initially to help us understand who was going to it and what they were uploading," said Jason Koler, Chief Information Security Officer at Eaton Corporation.

Agentic AI and the open-source DeepSeek model have opened new avenues for threat actors to exploit AI technologies, allowing them to automate and scale attacks at an unprecedented rate. The report notes that DeepSeek, originating from China, has begun to challenge established American players such as OpenAI, Anthropic, and Meta, providing strong performance, open access, and affordability, yet also introducing significant security challenges.

Enterprises provided substantial data volumes to AI tools, sending a total of 3,624 terabytes during the review period. This data movement signifies deep integration of AI into business operations. However, organisations blocked 59.9% of all AI/ML transactions, reflecting heightened awareness and proactive efforts to manage risks around data exposure, unauthorised access, and regulatory compliance.

"As AI transforms industries, it also creates new and unforeseen security challenges," said Deepen Desai, Chief Security Officer at Zscaler. "Data is the gold for AI innovation, but it must be handled securely. The Zscaler Zero Trust Exchange platform, powered by AI with over 500 trillion daily signals, provides real-time insights into threats, data, and access patterns—ensuring organisations can harness AI's transformative capabilities while mitigating its risks. Zero Trust Everywhere is the key to staying ahead in the rapidly evolving threat landscape as cybercriminals look to leverage AI in scaling their attacks."

Regionally, Australia has emerged among the top generators of AI/ML transactions, alongside the United States, India, Canada, Germany, Japan, and the United Kingdom. In the Asia-Pacific region, India led with 36.4% of activity, followed by Japan (15.2%) and Australia (13.6%). The global distribution saw the United States with 46.2% of transactions, followed by India (8.7%), the United Kingdom (4.2%), Germany (4.2%), Japan (3.6%), Canada (3.6%), and Australia (3.3%).

The finance and insurance sector generated the largest share of enterprise AI/ML traffic at 28.4%, with manufacturing following at 21.6%. The services (18.5%), technology (10.1%), healthcare (9.6%), and government (4.2%) sectors also showed substantial AI/ML activity, each encountering unique regulatory and security challenges amidst new AI-driven use cases such as fraud detection, risk modelling, supply chain optimisation, robotics automation, and customer service automation.

"The rapid rise of AI adoption across Australia and New Zealand is reshaping the way employees and organisations work, driving productivity and unlocking new possibilities. Industries like finance and manufacturing are leading the way, but this surge in AI usage also shines a spotlight on the urgent need for robust security measures to protect sensitive data and sustain innovation," said Eric Swift, Vice President & Managing Director, Zscaler Australia and New Zealand. "At Zscaler, we're seeing AI usage skyrocket—ThreatLabz has recorded a staggering 36-fold increase in AI transactions year-on-year globally. While this surge is helping businesses supercharge their operations, it also brings new cyber risks that we can't afford to ignore. The Zscaler Zero Trust Exchange is here to help businesses confidently embrace AI. With unmatched visibility, control, and security, we're ensuring that organisations in Australia and New Zealand can scale their AI adoption safely, boost innovation, and build trust in how sensitive information and data is handled."

The report indicates that, while the adoption of AI is delivering substantial productivity gains, it has also exposed organisations to a "rapidly evolving threat landscape". The need for upskilling is pronounced, with 83% of Australian business leaders prioritising AI adoption by 2025 and 40% identifying training as essential.

Zscaler continues to promote its zero trust security model as a measure to address these emerging risks. Key strategies detailed in the report include data classification, breach prediction, real-time AI insights, threat protection, and app segmentation, all designed to manage risk and limit exposure as enterprises increase their use of AI tools.