Report: 54% of Enterprises Hit by AI Agent Security Incidents
VentureBeat AI · rss · 2026-07-17
A VentureBeat survey of over a hundred mid-to-large enterprises reveals a stark "agent security gap": 54% of surveyed companies have experienced AI agent security incidents or near-misses (18% confirmed leaks, 36% proactively intercepted).
The core vulnerabilities lie in identity and access management:
- Rampant credential sharing: Only 32% of enterprises provision unique, restricted identities for each agent. A massive 69% share API keys or human account credentials among agents. Data shows that companies with credential sharing have a significantly higher incident rate (63.5%) compared to those enforcing strict unique identities (40.9%).
- Lack of isolation: Only 30% of enterprises run high-risk agents in sandboxes. Larger enterprises see higher incident rates, yet their adoption of sandbox isolation is paradoxically lower.
Defense tools and awareness:
- Corporate defenses rely heavily on native tools from major providers (like OpenAI guardrails or Microsoft cloud controls), with dedicated agent security tools almost entirely absent.
- Although average satisfaction with current security stacks sits high at 4.2/5, only a third believe their defenses outpace AI attackers, and most plan to replace their existing security tools within the year.
More from coding & agent
- Reticle builds a verification layer that lets AI coding agents check their own generated code — Scobleizer · 2026-10-27
- Trust your agent, but read its logs: Claude managed agents tips — brada · 2026-10-27
- Dev training a guard model to block prompt injection, tool and memory poisoning — gnukeith · 2026-10-27
- Dev trains a ~4GB-VRAM sentinel model to guard against prompt injection and poisoned memory — gnukeith · 2026-10-27
- Superset rebrands as Paradev, raises $11.5M seed led by USV — massimosgrelli · 2026-10-27
- Ex-Big Tech managers may thrive in the AI era: manage agents like you managed teams — lxfater · 2026-10-27