The annual rate of AI-related data incidents has risen sharply over the past several years, according to compiled incident tracking. Recorded incidents increased from roughly 3 per year in 2020 and 2021 to about 14 per year in 2025, a more than fourfold increase that tracks the broad adoption of AI tools across consumer and enterprise settings.

The growth reflects multiple categories of exposure. Reports through 2025 documented employees pasting confidential company information into public AI chatbots, creating leaks of proprietary data outside corporate controls. Separate incidents involved AI applications that stored user conversations with weak access controls, leaving private messages exposed to anyone who located the data.

High-profile cases illustrated the trend. In early 2026, a security firm's red-team exercise reported gaining read and write access to a production database behind a major consulting firm's internal AI chatbot within about two hours, reaching conversations on corporate strategy and client engagements before the vulnerability was patched. A separate leak at a popular consumer AI chat app exposed a large volume of user messages.

The pattern points to a widening gap between how quickly organizations deploy AI tools and how thoroughly they secure the data those tools collect. As adoption expands across business functions, security analysts expect the number of AI-related incidents to keep rising unless access controls, storage configuration, and data handling practices keep pace with deployment.

Source: StealthCloud - https://stealthcloud.ai/data/ai-privacy-incident-timeline/