Report finds 1,025% rise in AI vulnerabilities, many tied to APIs
Douglas Lopes via Unsplash
Wallarm has released its 2025 API ThreatStats Report, demonstrating that the prominent attack surface over the past year has been APIs. Furthermore, the biggest driver of API security risks was AI.
Researchers tracked and examined 439 AI-related CVEs, marking a 1,025% rise from the previous year. 99% of these were directly linked to APIs including including misconfigurations, injection flaws and new memory corruption vulnerabilities (resulting from AI’s dependence on high-performance binary APIs). Moreover, more than half of CISA’s recorded exploited vulnerabilities were API-connected for the first time, representing a 30% increase from the previous year.
The report found both legacy and modern APIs are the target of attacks. Legacy APIs are at risk due to outdated designs, while modern APIs are at risk due to improper configurations and integration difficulties.
CISA’s Known Exploited Vulnerabilities (KEV) Catalogue shows APIs are the largest category of leveraged vulnerabilities, with 33% being modern APIs. Exploits include injection attacks, improper authentication and API endpoint misconfigurations. 18% of leveraged vulnerabilities were legacy APIs in web applications. Common exploits include CSRF attacks, URL-based injection and outdated session-handling mechanisms.
Looking for a reprint of this article?
From high-res PDFs to custom plaques, order your copy today!
.webp?height=200&t=1692281098&width=200 "computer with multicolored text")
