AI-driven exposure outpaces visibility in expanding attack surfaces

At a glance

What this is: Infoblox says AI-driven attacks and expanding cloud, SaaS and AI attack surfaces are outpacing the value of visibility alone.

Why it matters: This matters because identity, access and exposure decisions now need to account for external digital assets, machine access paths and fast-changing AI-enabled threat activity.

By the numbers:

• Infoblox based the report on interviews with over 550 cybersecurity professionals.

Context

Attack surface expansion means the number of reachable assets, accounts, services and external dependencies grows faster than security teams can manually govern them. In practice, visibility tools can show what exists, but they do not by themselves decide what is truly exploitable, misowned or overexposed. That gap becomes sharper when cloud, SaaS and AI systems change constantly and create more places where access can be abused.

For identity and access governance, the important shift is that exposure management is no longer only a perimeter or asset inventory problem. It now intersects with NHI, delegated access, API connectivity and AI system trust boundaries, because each of those can create an externally reachable path into sensitive data or operational workflows.

Key questions

Q: What breaks when exposure management only measures visibility instead of risk reduction?

A: Teams end up with more findings but no better decision-making. Visibility without prioritisation creates backlog, duplicated effort, and stale risk status. The control failure is not discovery. It is the absence of a normalised model that can rank reachability, business context, and ownership so remediation work focuses on exposures that actually shrink attack paths.

Q: Why do cloud and SaaS identity paths increase exposure risk so quickly?

A: Because access scope, trust relationships, and configuration drift can combine into a routable attack path in minutes or hours. A temporary public route, over-permissioned role, or mis-scoped token can turn a small change into a viable compromise path before manual review catches up.

Q: How can security teams tell whether a fix actually reduced exposure?

A: They should test whether the original attack outcome can still be reproduced after remediation. If the same-scope retest cannot achieve the same impact, the fix has evidence behind it. If it can, the programme has only documented activity, not reduction.

Q: Should organisations treat AI systems like part of the attack surface or just another application?

A: They should treat AI systems as part of the attack surface whenever those systems have tokens, delegated access or connections to internal data and workflows. The governance question is not the model alone, but the identities and integrations that let it act on the environment.

Technical breakdown

Why visibility does not equal exploitable risk reduction

Exposure management is the discipline of discovering internet-facing and externally reachable assets, but it only becomes meaningful when tied to exploitability, ownership and control state. A dashboard can show that an asset exists without showing whether it is reachable through weak authentication, stale credentials, permissive API access or misconfigured cloud services. In modern estates, the problem is not lack of data. It is that security teams often lack a reliable way to prioritise what can actually be used against them across fast-changing SaaS, cloud and AI-connected environments.

Practical implication: Treat discovery as an input to risk triage, not as evidence that the exposure problem is solved.

How AI changes the attack surface and the control model

AI systems expand the attack surface in two ways. First, they introduce new externally reachable interfaces, data flows and trust relationships. Second, they accelerate attacker behaviour by enabling faster reconnaissance, content generation and social engineering at scale. That means the control model must cover not only the AI application itself, but also the identities, tokens, integrations and downstream services that let it act. Where AI is connected to business systems, machine identity governance becomes part of AI security, because access paths are now part of the threat boundary.

Practical implication: Map AI-to-system connections, tokens and delegated access as part of the same exposure programme.

Why preemptive security is replacing reactive cleanup

Preemptive security focuses on reducing real-world exploitability before an incident, rather than waiting to detect and contain abuse after exposure has already been discovered. That approach matters when attack surfaces change continuously and attacker dwell time can be short. In this context, the useful question is not whether a control produced more visibility, but whether it shortened the window in which an exposed asset, credential or integration could be used. The most mature programmes tie exposure findings to ownership, priority and remediation authority.

Practical implication: Use exploitability and business criticality to decide what gets fixed first, not raw volume of findings.

Threat narrative

Attacker objective: The attacker wants to turn publicly reachable exposure into usable access that can be leveraged for theft, disruption or persistence.

1. Entry occurs through an exposed cloud, SaaS or external digital asset that is visible to attackers and reachable from the internet or partner integrations.
2. Credential or access abuse follows when over-permissive tokens, delegated access or weakly governed identities let the attacker move from discovery to use.
3. Impact is realised as the exposed surface becomes a foothold for data access, workflow abuse or wider compromise across connected services.

Breaches seen in the wild

• Gravity SMTP CVE-2026-4020 API Keys Exposure: CVE-2026-4020 in Gravity SMTP exposes API keys via single HTTP request across 100,000 WordPress sites.
• Sisense breach 2024: A credential in Sisense's GitLab reportedly opened S3 buckets of customer tokens, passwords and certificates; CISA urged a full reset.

Read and download The State of NHI & AI Agent Breach Report 2026, covering 200+ breaches impacting Non-Human Identities including AI Agents.

NHI Mgmt Group analysis

Visibility is not governance: the market keeps treating exposure discovery as if it were the same thing as risk reduction. It is not. Discovery identifies what exists, but governance determines which exposed assets, identities and integrations are actually exploitable and who is accountable for fixing them. Practitioners should treat exposure programmes as decision systems, not scanning exercises.

AI is turning external reachability into a faster attack path: AI-driven threats compress the time between discovery and abuse, especially where cloud, SaaS and connected workflows share identities and tokens. That makes the exposure management problem inseparable from identity governance, because access paths are now part of the attack surface. Security leaders need to assume that reachable means targetable.

Exposure ownership is the new control point: the hardest problem is no longer finding assets, but assigning a durable owner for the decision to reduce risk. Without that, even high-fidelity visibility becomes a reporting layer rather than a control layer. The programme question is whether every exposed asset, credential and integration has a responsible team that can act.

Cloud, SaaS and AI converge on the same failure mode: overly broad access, weak lifecycle control and unclear trust boundaries. That convergence means practitioners should stop managing these domains as separate tool silos. The field is moving toward unified exposure and identity governance because exploitability crosses those boundaries.

Real-world exploitability must become the primary metric: a large inventory of findings is not the same as a reduced attack surface. Security programmes should measure whether exposure has been removed, constrained or made non-usable by an attacker. That shift is what separates a monitoring programme from a defensive one.

What this signals

Exposure programmes need an ownership model, not just a scanner. Security teams that cannot assign remediation authority to each exposed asset will keep producing findings faster than they reduce risk. The operational shift is from discovery volume to decision quality.

AI expands the governance perimeter as much as the technical one. Once AI systems can call internal services, the relevant control question becomes who owns their credentials, their trust boundaries and their downstream permissions. That is where NHI governance and AI security now overlap in practice.

For practitioners

• Prioritise exploitable exposures first Rank exposed assets by reachability, privilege and business criticality, then remediate the items that can be used immediately rather than the loudest findings.
• Govern delegated access paths Inventory SaaS, cloud and AI integrations that rely on tokens, webhooks, service accounts or other non-human credentials, and assign owners for lifecycle review.
• Constrain externally reachable identities Reduce standing privilege and remove unused tokens or accounts that create internet-facing paths into internal systems, especially where services interconnect across domains.
• Tie exposure findings to remediation authority Make sure every high-risk exposed asset has an accountable team that can approve containment, shutdown or hardening without waiting for periodic review cycles.
• Reassess AI-connected trust boundaries Review which AI systems can call internal services, what data they can reach and whether their credentials are governed with the same rigour as production workloads.

Key takeaways

• AI-driven threat activity is forcing exposure management to move beyond inventory and toward exploitability.
• Visibility improvements are useful only when they lead to faster decisions about what is reachable, over-permissive or externally exposed.
• Cloud, SaaS and AI systems now share the same governance problem, which is ownership of the identities and integrations that connect them.

Key terms

• Exposure management: Exposure management is the practice of identifying which assets are reachable by attackers and reducing that reach before exploitation occurs. For collaboration systems like SharePoint, it is not enough to know that a patch exists, because public accessibility changes the speed and likelihood of attack.
• Attack Surface Management: Attack surface management is the practice of finding and evaluating assets that could be exposed to misuse or compromise. CAASM focuses on internal visibility across the environment, while EASM focuses on externally reachable assets. It is a discovery discipline, not a complete identity control model.
• Verified Exploitability: Verified exploitability means a finding has been reproduced in execution, not merely inferred from code. It is the practical threshold that separates a plausible defect from a security issue that should drive severity, remediation priority, and incident response.
• Delegated Access: Delegated access is permission granted to one identity to act on behalf of another user, service, or system. In NHI environments, this usually appears in OAuth-connected apps and automation tooling. It is powerful, but it must be tightly scoped and reviewed because it can persist long after the original business need ends.

Deepen your knowledge

The NHI Foundation Level course, the industry's only accredited NHI security programme, covers NHI governance, machine identity security and secrets management. It is designed for practitioners who need to connect identity controls to operational security decisions.