TL;DR: AI security failures often emerge at the seams between lifecycle stages, according to Aqua Security. Lifecycle hand-offs can leave development, release, monitoring and governance disconnected, so ownership and tooling need to stay linked.
Editorial analysis by NHI Mgmt Group, based on content published by Aqua Security: “Navigating the OWASP LLM and GenAI Security Landscape”.
Key questions
Q: Why do lifecycle gaps create more risk than isolated AI security findings?
A: Because AI systems change across stages, a control that works in one phase may do nothing in the next.
Q: How can security teams tell whether extension controls are actually working for GenAI?
A: Look for runtime visibility into prompt mutation, injected queries, and result handling inside the browser, not just permission counts or URL blocks.
Q: What breaks when GenAI systems are secured like ordinary applications?
A: They miss the main risk surface.
Practitioner guidance
- Map AI controls to lifecycle stages Assign each GenAI security control to planning, development, testing, release, operation, monitoring, or governance so coverage gaps are visible.
- Review handoffs between control owners Identify where development, platform, security, and operations teams assume another team is covering the next lifecycle stage.
- Validate runtime controls against release assumptions Check whether deployment-time configuration checks and CI/CD validation still leave room for prompt injection, jailbreaks, or model misuse in production.
Bottom line: AI security failures often emerge at the seams between lifecycle stages, not inside a single control domain.
Glossary terms: Lifecycle Security, Shadow AI, Runtime Protection, Lifecycle Hand-off
Explore further
View Full Forum → | NHI Foundation Course → | Our Services → | Read the full analysis →
Lifecycle coverage is now the organising principle for GenAI security programmes. The article shows why AI security cannot be managed as a collection of separate safeguards bolted onto a model stack. Each stage of the lifecycle creates a different risk boundary, so governance has to follow the system from planning through operation. Practitioners should treat lifecycle mapping as the baseline for any credible AI security architecture.
A few things that frame the scale:
- Only 23% of IT leaders were very confident in their organisation's ability to manage security and governance for GenAI deployments, according to a 2025 Gartner survey of 360 IT leaders.
A question worth separating out:
Q: Should organisations prioritise discovery or access restriction first for shadow AI?
A: Discovery comes first, because teams cannot restrict what they cannot enumerate. Once agents, credentials, and data paths are visible, organisations can decide which permissions to remove, which workflows to register, and which use cases to ban. Restriction without discovery usually pushes the problem deeper into the environment.
👉 Read our full editorial: OWASP LLM and GenAI lifecycle mapping exposes AI security gaps