🔓 AI's Data Leak Problem is Growing

Our latest research found PII for over 90,000 people exposed by popular AI tools. Read the full report.

Hi there,   The rapid adoption of AI is creating a new, unfamiliar attack surface—and the data leak problem is growing faster than many realize.   Just as we saw in the early days of cloud computing, new technologies that prioritize rapid adoption often leave security as an afterthought. Our new research report, The State of AI Data Leaks 2025, provides a data-driven look at this emerging risk landscape. The report found that the risk is already deeply embedded in your supply chain, with 42% of top vendors using an AI data subprocessor like OpenAI or Anthropic. But what is the real-world leak rate of the technology these vendors are using?   Our research provides a first look at this critical issue, including insights about: Data leak statistics of popular AI tools, including ChromaDB's 35% exposure rate Leaked data, including PII for over 90,000 people, API keys, and business documents Actionable recommendations for assessing AI tools and educating employees on safe AI use History shows that awareness is the first step toward security. Get the data you need to protect your organization from this new generation of threats. Download the Full Report UpGuard Inc., 650 Castro Street, Suite 120-387, Mountain View, CA 94041, United States, +1 888 882 3223 Unsubscribe Manage preferences