ThreatLabz Report Finds Enterprise AI Use Surged as Data and Security Risks Grew
Summary
ThreatLabz’s analysis of nearly one trillion AI/ML transactions in 2025 found enterprise usage rose 83%, data transfers to AI tools increased 93%, and organizations blocked 39% of access attempts amid emerging security risks.
Key points
- AI/ML transactions grew 83% year over year across more than 3,400 applications observed in 2025.
- Data transfers to AI tools rose 93% year over year, increasing exposure of sensitive enterprise information.
- Organizations blocked 39% of AI/ML access attempts, while unsanctioned use and embedded SaaS AI features may have less oversight.
- ThreatLabz said every enterprise AI system tested failed at least once under realistic adversarial pressure.
- The report describes threat actors using AI to support social engineering, fake personas, evasion, and malware development.
- Finance and Insurance accounted for 23.3% of enterprise AI/ML activity, and Manufacturing for 19.5%.
Article Details
- Publisher
- Zscaler (ThreatLabz report)
- Report Period
- 2025
- Scope
- Enterprise AI/ML activity and security across global enterprises, based on transactions observed through the Zscaler Zero Trust Exchange.
- Sample Size
- Nearly 1 trillion AI/ML transactions across more than 3,400 applications; the report also describes red-team testing of enterprise AI systems without specifying the number tested.
- Key Statistics
- Enterprise AI/ML transactions increased 83% year over year in 2025; analysis covered more than 3,400 applications generating AI/ML traffic.
- Data transfers to AI tools increased 93% year over year in 2025, reaching tens of thousands of terabytes in total.
- Enterprises blocked 39% of AI/ML access attempts in 2025.
- Finance & Insurance accounted for 23.3% of enterprise AI/ML activity; Manufacturing accounted for 19.5%.
- Every enterprise AI system tested in ThreatLabz red-team testing failed at least once under realistic adversarial pressure; the number of systems tested was not stated.
- Recommendations
- Maintain visibility into AI usage, exposure, and dependencies across applications, models, pipelines, and supporting infrastructure, including through AI bills of material.
- Use granular access controls and inline inspection of prompts and responses to prevent sensitive data from being sent to external models and unsafe outputs from being returned.
- Protect enterprise AI applications and infrastructure with system hardening, runtime protections, vulnerability detection, and adversarial red-team testing.
- Align AI controls with industry risk profiles, compliance requirements, and operational dependencies.
- Explicitly classify and monitor embedded AI features and their data flows within approved SaaS applications.
Vendors
Products
ChatGPTWhen analyzing AI/ML activity across the full year, the most widely used applications were Grammarly, ChatGPT, and Microsoft Copilot, reflecting how deeply AI is now embedded in everyday work. Codeium also rankedCodeiumGrammarly, ChatGPT, and Microsoft Copilot, reflecting how deeply AI is now embedded in everyday work. Codeium also ranked among the top applications by transaction volume, underscoring the growing role of AI inGoogle GeminiWhile the enterprise AI adoption landscape continues to evolve, with models such as Google Gemini and Anthropic gaining traction more recently, enterprise usage in 2025 remained concentrated around a small set ofGrammarlytools. When analyzing AI/ML activity across the full year, the most widely used applications were Grammarly, ChatGPT, and Microsoft Copilot, reflecting how deeply AI is now embedded in everyday work. Codeium alsoMicrosoft CopilotAI/ML activity across the full year, the most widely used applications were Grammarly, ChatGPT, and Microsoft Copilot, reflecting how deeply AI is now embedded in everyday work. Codeium also ranked among the topZscaler Zero Trust ExchangeThe report draws on analysis of nearly one trillion AI and ML transactions observed across the Zscaler Zero Trust Exchange™ throughout 2025. That activity translates to hundreds of thousands of AI transactions per
Industries
Finance & Insurancewas not uniform. Each sector is moving at a different pace and with different levels of oversight. Finance & Insurance once again generated the largest share (23.3%) of enterprise AI/ML activity. Manufacturing remainedManufacturingFinance & Insurance once again generated the largest share (23.3%) of enterprise AI/ML activity. Manufacturing remained highly active at 19.5%, driven by automation, analytics, and operational workflows.