Threat Labs Report: Asia 2025

GenAI adoption is rapidly accelerating across Asia, with 93% of organizations now using genAI applications, up from 84% just a year ago. In the region, 98% of organizations use tools that incorporate genAI-powered features, and 98% rely on applications that leverage user data for training. Overall, 92% of organizations in Asia have adopted genAI apps directly, placing regional adoption slightly above global trends, where 90% of organizations have embraced cloud-based genAI solutions.

While overall adoption has grown, the use of personal genAI accounts remained largely consistent at around 79% through December 2024, before showing a sharp decline from January 2025 onward, declining to 35% by September 2025. In contrast, adoption of organization-approved genAI solutions rose significantly during the same period, increasing from 17% to 55%. This is a very promising trend, showing that most organizations have been successful in migrating users away from personal genAI apps where the organization has little control over data security, to enterprise genAI solutions that provide more robust controls and auditing capabilities. Organizations that have not yet made the switch should look to this trend as inspiration to prioritize secure, enterprise-grade genAI solutions that enable innovation while ensuring compliance and reducing the risk.

In Asia, the top 10 most widely used genAI applications closely reflect global adoption patterns. ChatGPT leads the list, used by 75% of organizations, while Google Gemini follows closely at 74%, making the two platforms nearly equal in popularity across the region. Microsoft Copilot has been adopted by 50% of organizations, with Microsoft 365 Copilot at 47%, underscoring rising demand for genAI solutions integrated directly into productivity platforms. The remainder of the top 10 is made up of a diverse mix of domain-specific applications and embedded AI tools tailored to business needs and regional use cases.

The chart below illustrates how the popularity of the top 10 genAI applications has evolved across Asia over the past year, highlighting the rapid pace of change in the regional landscape. ChatGPT usage has remained largely constant, maintaining its position as the most widely used app. In contrast, Google Gemini has shown strong momentum, rising from 45% in September 2024 to 74% by September 2025, reflecting accelerating adoption across the region. Both Microsoft Copilot and Microsoft 365 Copilot recorded notable growth beginning between late 2024 and early 2025, driven by their integration into widely used productivity platforms. Notably, Grok has entered the top 10 for the first time, signaling the emergence of new competitors in Asia’s genAI ecosystem.

As genAI adoption accelerates across enterprises in Asia, the risks tied to data exposure are becoming a critical concern. Organizations in the region are increasingly relying on genAI tools for tasks such as summarizing large documents, generating content, and streamlining software development workflows. However, these use cases require users to share potentially sensitive data with genAI applications, expanding the surface for data security threats.

Data security remains a pressing issue as genAI becomes embedded in daily business operations, with risks around data loss growing in relevance. The rise of shadow AI further compounds the challenge, as employees continue to adopt tools outside of official approval channels. Security teams can’t secure what they do not see, and as AI usage becomes increasingly sophisticated and widespread, it is critical that they prioritize gaining visibility of individuals experimenting with AI within their organization to ensure sensitive data is not exposed through these deployments and usage.

Analysis of recent data policy violations shows that the most common type of sensitive information exposed to genAI apps in Asia is source code, accounting for 62% of all incidents. Regulated data makes up 18% of violations, while intellectual property represents 14%. Other exposures, including passwords and API keys, are also present, though their frequency is lower than the global trend.

Organizations across Asia are taking a cautious approach to genAI tools, with many blocking specific applications due to security, privacy, or compliance concerns. While policies vary by company, some apps are restricted far more often than others, highlighting where organizations see the greatest risks. If any of the apps listed below are present in your environment, consider reviewing their usage and assessing whether applying restrictions at the category level may offer stronger protection than managing each tool individually

In the region, DeepSeek tops the list of most blocked genAI applications, with 46% of organizations restricting access. Concerns primarily center on transparency and the risks tied to rapidly emerging genAI platforms. ZeroGPT follows closely at 43%, primarily due to the way it handles user data, including storing submitted content and reports of redirecting information to third-party websites.

These patterns suggest that organizations in Asia are not only reacting to risks tied to specific applications, but also reinforcing their broader strategies for managing genAI within established risk and compliance frameworks.