AI-related API vulnerabilities have increased by 1025% year over year, underscoring the escalating risks associated with autonomous agentic AI systems.
Wallarm’s ThreatStats report for Q1 2025, titled “The Rise of Agentic AI,” highlights that evolving API threats are fueled by the rise of agentic AI systems, growing complexity in cloud-native infrastructure, and a surge in software supply chain risks.
“In the first quarter of 2025, overall API threats continued to increase across multiple industries, from healthcare to AI and beyond,” said Ivan Novikov, CEO and Co-Founder of Wallarm. “Our research shows that AI agent security risk largely stems from APIs.”
Agentic AI Exacerbates Familiar Security Risks
Wallarm’s analysis of 2869 agentic AI security issues on GitHub revealed that 65% were API-related, reflecting the intrinsic link between agentic AI security and API security.
The Common Weakness Enumeration (CWE) data reflects similar patterns. While the most frequent security issue is the use of Unmaintained Third-Party Components (CWE-937)—a common software weakness—the second and third places were Improper Input Validation (CWE-20) and Uncontrolled Resource Consumption (CWE-400), both of which are associated with API vulnerabilities.
Misconfigurations Dominate API Breaches
The report also reviews the most significant API-related breaches from Q1 2025, highlighting ongoing issues with misconfigurations and insufficient access control. The top five incidents were:
- Oracle Cloud: An attack exploiting an unpatched CVE-2021-35587 vulnerability exposed 6 million records, demonstrating the long-term risks of legacy systems.
- Deepseek: Missing authentication resulted in the exposure of over 1 million records.
- Common Crawl: A public dataset used to train large language models contained 11,908 live API secrets, exposing the widespread risks in the software supply chain.
- Volkswagen: Weak JSON Web Token (JWT) implementations led to a breach exposing 800,000 records.
- NHS UK: An unauthenticated API endpoint exposed extensive patient data, highlighting vulnerabilities in healthcare sector infrastructure.
According to Wallarm, breaches tied to misconfiguration, hardcoded secrets, and unauthenticated API access dominated the quarter, with AI and healthcare organizations disproportionately affected.
Access Control Failures are a Growing Concern
Wallarm’s analysis also found that three major agentic AI vulnerabilities – CWE-285 (Improper Authorization), CWE-284 (Improper Access Control), and CWE-287 (Improper Authentication) – were directly related to access control flaws. In total, 209 CVEs recorded during the quarter fell into the Broken Access Control category, making it the largest in Wallarm’s vulnerability dataset.
Recommendations
In response to these findings, Wallarm recommends implementing:
Threat Management:
- Integrating API-specific threat intelligence.
- Refreshing API threat models quarterly.
Real-Time Protection:
- Monitoring API traffic and blocking anomalies in real time.
- Prioritizing investment in API security tools and training, particularly for real-time defense, secret scanning, and schema validation.
Agentic AI Security:
- Expanding API discovery to include agentic AI endpoints.
- Implementing dedicated agentic AI security strategies, including stricter access controls and operational guardrails.
Josh is a Content writer at Bora. He graduated with a degree in Journalism in 2021 and has a background in cybersecurity PR. He's written on a wide range of topics, from AI to Zero Trust, and is particularly interested in the impacts of cybersecurity on the wider economy.
The opinions expressed in this post belong to the individual contributors and do not necessarily reflect the views of Information Security Buzz.