Safetica Research Finds AI Is Now the No. 1 Risky Application Category for Data Protection

SAN JOSE, California, September 3, 2026 — Safetica, a global leader in Intelligent Data Security, today released its H1 2026 Data Protection Trends Report. The research shows sensitive data risk is occurring across AI tools, cloud platforms, and everyday communication channels.

Safetica's Research Team analyzed hundreds of thousands of blocked activities, data policy violations, unusual data-handling events, and risky application interactions across Safetica customer environments worldwide during the first half of 2026.

AI stood out in the data. AI tools became the riskiest application category in Q2, accounting for 40.7% of flags. ChatGPT accounted for 44.8% of blocked AI activity, and Microsoft Copilot accounted for 37.5%. Together, the two platforms represented 82.3% of blocked AI activity.

The report also shows a broader shift beyond AI. Data risk is increasingly concentrated in the platforms and workflows people use to get work done.

Four data protection trends defined H1 2026

1. Core platforms are the top risk surface. Google and Microsoft sites accounted for 43.7% of all blocked activity in Q2, nearly triple their combined 14.2% share in Q3 2025.
2. Everyday tools still carry most of the risk. Email, web and instant messaging accounted for more than 72% of data policy violation paths in Q2. Presentations and text files represented more than 57% of file types involved.
3. The spotlight on AI tool risk continues. AI represented 40.7% of risky application flags in Q2, while AI tools' share of overall blocked activity increased from 1.4% in Q3 2025 to 9.9% in Q2 2026.
4. Insider risk migrated to the network. Network-triggered Dynamic DLP activity increased from 5.1% in Q1 to 21.6% in Q2, while network-related Unusual Activity flags also increased sharply.
   

“The way people work is changing, and data protection has to change with it. AI is becoming part of everyday work alongside email, collaboration platforms and cloud applications. Security teams need to understand how sensitive data is being used across those interactions so they can protect it without unnecessarily disrupting how people work.”

Mirek Kren, CEO, Safetica

AI is changing the interaction, not the underlying data protection problem

Sensitive data risk is appearing in widely adopted productivity platforms, communication tools, and AI services.

Security teams need visibility into data, users, applications, and actions to distinguish routine work from potential exposure.

The H1 findings also continue trends identified in Safetica's H2 2025 research, which found data exposure increasingly embedded in routine workflows including productivity suites, web applications, messaging, AI tools, screenshots, and removable media.

Download the H1 2026 Data Protection Trends Report

The complete report includes quarter-over-quarter analysis of blocked activity, AI platforms, policy violations, unusual user behavior, Dynamic DLP triggers, file types, and industry-specific trends.

Download the report: solutions.safetica.com/h1-2026-data-protection-trends-key-findings

About Safetica

Safetica is a global leader in Intelligent Data Security. Our AI-powered platform unifies data protection, insider risk management, compliance readiness, data discovery, and AI Protection across cloud and on-premises environments.

Safetica helps security teams protect sensitive data wherever it moves, and however it is used, including through AI, while reducing insider risk and supporting compliance with minimal disruption to business workflows.