AI use in businesses jumps 3,000 percent as cyber threats grow

A new report shows that more companies are using artificial intelligence (AI) tools than ever before. According to Zscaler’s 2025 AI Security Report, enterprise AI usage grew by more than 3,000 percent between February and December 2024. But this growth has also brought a rise in cybersecurity risks.

Zscaler’s ThreatLabz team studied over 536 billion AI-related transactions on its Zero Trust Exchange platform. They found that many businesses face real threats such as AI-powered phishing, fake AI platforms, and dangers from open-source tools like DeepSeek.

ChatGPT leads both usage and security concerns

The report found that ChatGPT accounted for 45.2 percent of all AI and machine learning activity. It was also the most blocked AI tool by businesses. Grammarly and Microsoft Copilot were the next most-blocked platforms.

Many businesses are worried that employees may share sensitive information using these tools, often without proper approval. This has led companies to take steps to monitor and limit AI use.

DeepSeek and agentic AI pose new risks

Agentic AI refers to systems that can take actions on their own. DeepSeek, an open-source AI model from China, is a growing concern. It offers high performance and low cost, but it also makes it easier for cybercriminals to launch advanced attacks.

The report warns that such tools can be used to automate and scale attacks faster than ever before.

Data shared with AI tools is massive

Businesses sent a total of 3,624 terabytes of data to AI platforms during the study period. This shows how deeply AI is now built into daily operations. But companies also blocked 59.9 percent of AI and ML transactions. This reflects growing awareness of data risks and the need for strong controls.

Deepen Desai, Chief Security Officer at Zscaler, explained that data is the most important resource for AI. But it must be protected to avoid harm. Zscaler’s platform uses over 500 trillion daily signals to detect threats and ensure secure access.

AI usage varies by country and region

The United States had the highest share of global AI transactions at 46.2 percent. India followed with 8.7 percent, then the United Kingdom and Germany at 4.2 percent each. Japan and Canada both had 3.6 percent, while Australia had 3.3 percent.

In the Asia-Pacific region, India led with 36.4 percent of activity. Japan was next with 15.2 percent and Australia had 13.6 percent.

Key industries using AI tools

The finance and insurance sector led all industries with 28.4 percent of AI traffic. Manufacturing followed with 21.6 percent. Other sectors using AI heavily included:

Services at 18.5 percent
Technology at 10.1 percent
Healthcare at 9.6 percent
Government at 4.2 percent

These industries use AI for many tasks including fraud detection, risk analysis, supply chain management, robotics, and customer support.

AI changing work in Australia and New Zealand

Eric Swift, Zscaler’s Vice President for Australia and New Zealand, said the rise of AI is reshaping how people work. He noted a 36-fold increase in AI use year-over-year. While this is helping companies become more efficient, it also brings serious cyber risks.

He emphasized the need for strong security to support continued innovation and protect sensitive information.

Workers need training to keep up with AI

The report found that 83 percent of business leaders in Australia aim to adopt AI by 2025. But 40 percent said that training workers is essential to doing this safely.

Zscaler suggests a Zero Trust model to manage these risks. This approach checks every request to access data or systems, whether inside or outside the company network.

Ways to stay secure with AI

To reduce cyber risks, companies are advised to follow these steps

Use data classification to manage sensitive files
Apply real-time AI monitoring to detect unusual behavior
Segment apps to control who can access what
Strengthen access rules for AI platforms
Train employees on how to use AI safely

These actions can help companies stay protected while still taking advantage of AI tools.