Developer @0x_Osprey has launched the CVE Tracker website, a tool designed to visualize the growth in Common Vulnerabilities and Exposures (CVE) reports. The data shows a sharp increase in the number of identified security vulnerabilities starting from 2026.
The platform allows users to overlay the release timelines of major AI models with vulnerability discovery peaks. According to reports, AI companies frequently conduct vulnerability research using test versions of their models before the official public launch, which may cause discrepancies between release dates and spikes in reported flaws.
AI is increasingly acting as a catalyst in security research, enabling faster code scanning, anomaly reproduction, and report generation. This automation allows researchers to identify issues in software areas that were previously difficult to cover manually.
However, the surge in reports does not necessarily indicate that systems have become more dangerous. The influx of data includes potential risks alongside a high volume of low-quality or duplicate reports, which increases the workload for open-source maintainers and security teams responsible for triaging these findings.
