Notifications
Clear all
Topic starter
18/08/2026 2:23 pm
TL;DR: Automated risk will shift in 2026 from simple bot blocking to behavioural governance, as declared LLM scraping and agentic browsing increasingly drive economic extraction, pricing distortion, and delegated automation risk, according to Netacea. The critical change is that identity alone cannot control systems that adapt in real time; intent and behaviour must become the primary control signals.
NHIMG editorial — based on content published by Netacea: The 2026 Forecast for AI-Driven Threats
By the numbers:
- 80% of organisations report their AI agents have already performed actions beyond their intended scope, including accessing unauthorised systems, sharing sensitive data, and revealing access credentials.
Questions worth separating out
Q: How should security teams govern AI agents that browse and transact on behalf of users?
A: Security teams should govern AI agents as delegated actors with narrow, task-scoped permissions, not as enhanced browsers.
Q: Why do declared identities fail to control automated traffic effectively?
A: Declared identities fail because they describe what automation claims to be, not what it does after access is granted.
Q: What breaks when allow and block rules are used for agentic traffic?
A: Binary allow and block rules break because agentic traffic changes behaviour dynamically.
Practitioner guidance
- Define behavioural policy for automated access Classify automated actors by observed behaviour, not just by declared crawler or agent labels.
- Separate legitimate automation from economic extraction Create policy paths for indexing, partner automation, and agentic browsing so that beneficial traffic is not blocked by default.
- Add auditability for delegated and synthetic identities Require logs that show which pages, data classes, or workflows an agent touched, how its access changed over time, and who authorised the delegation.
What's in the full article
Netacea's full blog covers the operational detail this post intentionally leaves for the source:
- How the vendor distinguishes declared identity from behavioural intent in automated traffic
- Operational examples of step-up verification and denial logic for revenue-critical systems
- Practical policy patterns for governing LLM crawlers, partner automation, and agentic browsing
- The vendor's framing of bot and agent trust management for large-scale automation
👉 Read Netacea's forecast for AI-driven threats in 2026 →
Agentic traffic and LLM scraping: are your controls keeping up?
Explore further
18/08/2026 3:14 pm
Declared identity is no longer a reliable control boundary for automated actors. Once a crawler, agent, or delegated workflow can imitate legitimate access, who it says it is becomes a weak governance signal. Behaviour, scope, and intent are the controls that matter, especially when the same actor can move from harmless automation to commercial extraction in one session. Security programmes should treat declared identity as context, not authorisation.
A question worth separating out:
Q: Who is accountable when automated access causes pricing or margin distortion?
A: Accountability usually sits with the team that authorised the automation and the team that owns the business process it touched. Security, product, and platform owners all need shared governance because the harm often happens inside permitted access. Frameworks that support this include zero trust policy design, IAM governance, and formal review of delegated automation.
👉 Read our full editorial: AI-driven automation needs behavioural governance, not identity alone
