AI identity sprawl is widening faster than IAM governance

At a glance

What this is: This report says AI adoption is nearly universal, but the identities AI creates are spreading faster than IAM governance can see or control them.

Why it matters: For IAM, IGA, PAM, and NHI teams, the issue is that AI tools are becoming privileged actors with weak inventory, review, and offboarding discipline.

By the numbers:

• 99.6% of organizations are already using or actively planning to adopt artificial intelligence.
• 42% of IT leaders are already concerned about a lack of identity governance for non-human actors.
• Only 23% of IT teams are actively securing agentic AI identities.
• 51% of leaders cite AI tools integrating with sensitive systems without proper review as a top risk.

Context

AI identity sprawl is the expansion of non-human identities created by AI tools, bots, scripts, API keys, and autonomous agents across systems that were designed for human users. In this report, the governance gap is not adoption itself but the speed at which machine identities appear outside normal IAM visibility and control.

JumpCloud's article argues that traditional IAM models do not keep pace when AI tools connect to sensitive systems without review. That matters because the identity problem is no longer just who uses AI, but what identities AI creates and what privileges those identities inherit.

The article's starting point is typical of current enterprise conditions. AI adoption is broad, but governance over the resulting non-human identity estate remains immature and uneven.

Key questions

Q: What breaks when AI identities are handled outside IAM?

A: When AI identities sit outside IAM, organisations lose a consistent record of who has access, why access exists, and who approved it. That creates policy drift, weak accountability, and incomplete audit evidence. The programme may still function operationally, but it will not provide dependable governance or defensible compliance evidence.

Q: Why do AI systems increase identity risk even when they improve security operations?

A: AI can help defenders, but it also helps attackers scale phishing, impersonation, and credential abuse. That means the same adoption that improves detection can also widen exposure unless authentication, monitoring, and access governance keep pace.

Q: How do security teams respond when AI identity governance is already deficient?

A: First, contain the highest-risk identities by reviewing standing access, removing unnecessary privileges, and forcing ownership assignment for every NHI. Then establish discovery and certification workflows so the same problem does not reappear. If AI is already in production, the right response is staged reduction of exposure, not a blanket freeze on adoption.

Q: Why do organisations need to control non-human identities separately from human users in AI-enabled workflows?

A: Non-human identities often outnumber human users and operate at machine speed, so their risk profile is very different. They can hold persistent access, call APIs directly, and trigger actions without a human present. Separate governance reduces over-privilege, improves traceability, and prevents automated systems from becoming a hidden attack path.

Technical breakdown

Why AI tools become non-human identities

AI tools are not just applications consuming services. When they act through scripts, bots, API keys, or autonomous agents, they become non-human identities that can authenticate, query, and execute work across systems. The technical issue is that these identities often sit outside the user directory and outside the normal joiner-mover-leaver model, so they evade the governance workflows built for people. Once they are embedded in code, orchestration, or platform settings, they are easy to miss and difficult to enumerate consistently across environments.

Practical implication: Map AI-enabled execution paths to NHI inventory processes, not to human user administration.

Why elevated privileges make agentic AI harder to govern

Agentic AI often needs broad privileges to complete tasks, which creates a mismatch between operational usefulness and governance precision. Least privilege is harder to define when the system can initiate actions, call tools, and interact with multiple services without a human deciding each step. That is a structural challenge for IAM because authorisation was historically assigned to predictable human roles, not to systems whose task scope expands or changes at runtime. The more sensitive the target systems, the more damaging any privilege drift becomes.

Practical implication: Treat AI-connected access as a high-risk privilege domain and scope it from issuance time, not after deployment.

Why invisible identities defeat review and offboarding

The report's core governance warning is that these identities are often hidden in code, API configuration, and service integrations rather than in standard identity records. That means access review, certification, and offboarding can miss them even when teams believe controls are working. If the identity is not visible, it cannot be recertified, and if it is not owned, it cannot be reliably revoked. This is why AI governance must track the identity itself, not only the application that uses it.

Practical implication: Require ownership, inventory, and revocation paths for every AI-created identity before it reaches production.

Threat narrative

Attacker objective: The objective is to exploit ungoverned AI-linked identities as privileged access paths into critical enterprise systems.

1. Entry occurs when AI tools are connected to sensitive systems through scripts, bots, API keys, or autonomous agents that sit outside standard user governance.
2. Credential access and privilege use follow when those identities inherit elevated permissions needed to function across databases, applications, or privileged accounts.
3. Impact emerges when an unreviewed AI identity becomes a hidden backdoor into critical systems and expands the organization's attack surface.

Breaches seen in the wild

• JumpCloud breach 2023: North Korean hackers breached JumpCloud and abused its device commands framework against a few customers; all admin API keys were reset.

Read and download The State of NHI & AI Agent Breach Report 2026, covering 150+ breaches impacting Non-Human Identities including AI Agents.

NHI Mgmt Group analysis

AI identity sprawl is now an identity governance problem, not an AI adoption side effect. The article shows that AI is spreading into core business workflows faster than identity teams can inventory the actors it creates. That means the control plane has to shift from human-centric account management to lifecycle governance for the new machine identities AI introduces.

Invisible AI identities create a governance gap that traditional IAM cannot close on its own. When identities are buried in code, API settings, and service integrations, they do not participate in the normal directory-based controls that many programmes still treat as authoritative. The result is not just missing coverage but a false sense of completeness, which is exactly where attack surface grows unnoticed.

AI tools integrating with sensitive systems without proper review is the right named concept for this risk. The article's 51% figure shows that practitioners are already identifying the failure mode, but identification is not control. Once sensitive-system access is embedded before review, IAM becomes reactive rather than governing. The practitioner conclusion is that review must occur at identity creation, not after integration.

Unified identity governance is becoming the minimum viable operating model for human and machine access. The report's direction is clear: separate governance lanes for human users and AI-created identities leave gaps in ownership, certification, and offboarding. A single programme view does not mean a single control set, but it does mean one accountable inventory of who and what can act.

Shadow AI will continue to expand wherever security teams cannot govern the identity layer. The article links avoidance and workarounds to unmanaged adoption, which means blocking AI without providing a governance path is self-defeating. The field implication is that the identity programme has to become an enabler of controlled AI use, or it will be bypassed.

From our research library:

• Organisations that describe themselves as confident in their AI deployment actually experience a 72% security incident rate, compared to 33% for those who remain cautious, according to the 2026 Infrastructure Identity Survey.
• Only 13% of organisations feel extremely prepared for the reality of agentic AI despite the majority racing toward autonomous adoption, according to the 2026 Infrastructure Identity Survey.
• Read next: Agentic AI Identity Guide

What this signals

AI identity sprawl is the control problem hiding inside AI adoption. Organisations are adding AI faster than they are adding ownership, inventory, and revocation discipline for the identities those systems create. That means the practical question is no longer whether AI is present, but whether the identity programme can see and govern the actors behind it.

Agentic AI identity governance will increasingly sit alongside workload identity and secrets management. The same operational pattern keeps repeating: credentials are created in one layer, consumed in another, and never fully re-certified. Teams that already manage service accounts and API keys should treat AI-connected identities as the next expansion of that same governance surface.

For practitioners

• Inventory AI-created identities separately from human accounts Catalogue scripts, bots, API keys, service credentials, and autonomous agents that access production systems, then assign ownership for each identity.
• Review AI-to-system integrations before production use Require explicit approval for any AI tool that reaches sensitive databases, privileged accounts, or critical applications, with access scope documented and versioned.
• Apply lifecycle controls to non-human AI identities Define joiner, mover, and leaver handling for AI identities so creation, change, and offboarding are managed with the same discipline as other high-risk access.
• Restrict privilege where AI tools can execute tasks Use scoped credentials, separated environments, and explicit revocation paths so AI systems cannot retain broad standing access after their task changes.
• Build governance metrics for shadow AI exposure Track how many AI-connected identities exist, who owns them, and how many have never been reviewed or recertified, then report that gap to risk owners.

Key takeaways

• AI adoption is broad, but the identities created by AI tools are spreading faster than traditional IAM can inventory and control.
• The main failure is governance visibility, because machine identities hidden in code and integration settings evade review and offboarding.
• Practitioners need one identity programme that treats AI-created credentials as governed assets, not as incidental configuration.

Key terms

• AI Identity Sprawl: AI identity sprawl is the uncontrolled growth of identities created for AI systems, agents, tools, and automation. It occurs when service accounts, API keys, tokens, certificates, and delegated permissions multiply across environments without clear ownership, lifecycle control, or review. This increases exposure, weakens accountability, and makes governance difficult.
• Non-Human Identity (NHI): A digital identity assigned to a non-human entity such as a software application, service account, API key, bot, machine, or AI agent that enables it to authenticate and interact with systems without direct human involvement. NHIs now outnumber human identities in most enterprises by 25 to 50 times.
• Shadow AI: AI agents, copilots, or connected tools operating without full visibility or governance from security teams. Shadow AI becomes an identity problem when those systems authenticate with unmanaged tokens, service accounts, or OAuth apps that can reach production resources.
• Agentic AI Identity: The complete set of credentials, permissions, and governance controls applied to an autonomous AI agent, covering authentication, authorisation, action logging, and access revocation. Distinct from traditional NHI because agent identities are often ephemeral, delegated, and multi-hop.

Deepen your knowledge

NHI governance, agentic AI identity, and machine identity lifecycle are core topics in our NHI Foundation Level course, the industry's only accredited NHI security programme. If you are building or maturing an IAM programme, it is worth exploring.