As AI adoption expands across workplace learning, one question is becoming increasingly difficult for organizations to answer clearly: Who actually owns AI in learning? 

At first glance, the answer appears relatively straightforward. Learning teams introduce AI to accelerate content development, improve learner support, enable simulations, or reduce operational bottlenecks. But as AI systems become more deeply embedded within enterprise environments, ownership quickly becomes more complicated. Questions emerge around infrastructure governance, data privacy, vendor accountability, ethical oversight, integration architecture, and operational responsibility. At that point, AI no longer sits comfortably within the boundaries of a single function.

Across many of the organizations we have been studying, this is creating a new governance tension. L&D often owns the learning experience, but IT owns infrastructure. Security governs risk exposure. Legal evaluates compliance and liability. Procurement manages vendor relationships. HR oversees workforce implications. Meanwhile, AI systems increasingly operate across all of these domains simultaneously, making ownership both distributed and ambiguous.

This article reflects on the governance problem emerging inside enterprise learning and argues that AI adoption is forcing organizations to rethink ownership itself. The challenge is no longer simply who manages the tool, but how organizations create governance structures capable of supporting systems that operate across multiple functions, workflows, and layers of accountability at once.

This article draws on ongoing research by CommLab India in collaboration with Lancaster University, exploring how AI is shaping workplace learning across large organizations. The observations here are based on anonymized conversations with learning leaders navigating these changes in practice.

For many years, ownership inside workplace learning was relatively easy to define. L&D owned the learning strategy and experience. IT managed infrastructure and technical support. Procurement handled vendors and contracts. Compliance reviewed regulated content when required. The boundaries between functions were not always perfect, but they were generally understandable and operationally stable.

AI changes that stability — and the numbers are beginning to reflect the scale of the problem. According to the IAPP's AI Governance Profession Report 2025, 77% of organizations are actively working on AI governance, with that figure rising to nearly 90% for organizations already using AI. Yet despite this widespread recognition, formal governance structures remain the exception rather than the rule.

Because AI systems do not behave like traditional learning technologies. They do not simply host content or automate administrative workflows. They generate outputs, interact with enterprise data, influence decisions, adapt dynamically, and increasingly integrate into broader organizational systems that extend well beyond the learning function itself.

The moment this happens, ownership becomes far more difficult to define clearly. The questions that surface are not merely technical — they are governance questions:

And across many organizations, the governance structures required to answer them are still evolving in real time.

One of the clearest patterns emerging across enterprise AI adoption is that AI systems rarely fit neatly into existing organizational structures. As ISACA has observed, what makes AI governance uniquely complex is its intersectional risk profile — where privacy, cybersecurity, and regulatory compliance converge in unprecedented ways, making siloed approaches fundamentally inadequate.

This structural tension is not abstract.

L&D often feels responsible for learning quality, learner experience, instructional effectiveness, and capability development.

At the same time,

Meanwhile, AI vendors themselves increasingly shape how systems behave through model updates, evolving capabilities, and shifting data governance practices.

"Responsible AI must be governed in an integrated and transparent fashion across the enterprise. This can only occur through shared accountability, anticipation of risk, and the building of trust through actions rather than assumptions."

— ISACA, Collaboration and the New Triad of AI Governance, 2025

As a result, ownership becomes distributed across multiple functions. But distributed ownership can easily become unclear ownership — and unclear ownership creates governance instability.

Why AI Creates Governance Complexity in Learning

One of the most visible governance tensions emerging inside organizations is the evolving relationship between IT and L&D. Traditionally, learning teams selected platforms while IT ensured compatibility, infrastructure support, and security compliance. The relationship was important, but the boundaries were relatively clear.

AI changes the nature of the systems being introduced. AI-enabled learning tools increasingly require enterprise integration, API governance, identity and access management, infrastructure scalability, data protection oversight, and platform interoperability — all of which naturally expand IT involvement. At the same time, learning teams remain closest to the actual operational use cases: simulations, coaching systems, AI-supported content generation, and learner support environments.

Gartner's 2025 Market Guide for AI Governance Platforms confirms this shift, noting that the rapid adoption of generative and agentic AI is outpacing traditional governance structures — with organizations struggling with distributed oversight and rising regulatory pressure. The guide signals that AI governance platforms have become essential infrastructure for managing AI at scale, suggesting that neither IT nor L&D can govern these systems independently.

Gartner VP Analyst Sumit Agarwal has noted that traditional AI governance models built on periodic audits and static policies cannot keep up with nondeterministic, modern AI architectures — requiring governance mechanisms embedded directly into the AI architecture itself. This fundamentally changes the relationship between learning teams and the technical functions that support them.

Another major governance issue emerging across organizations is the growing importance of vendor oversight. In traditional learning systems, vendors primarily supplied platforms or infrastructure with relatively stable functionality. AI vendors operate differently — they continuously influence how outputs are generated, how models evolve, how organizational data is processed, and what capabilities become available.

This introduces a much deeper dependency relationship. Organizations are no longer simply purchasing software — they are entering into ongoing relationships with systems that continue to evolve long after implementation.

PwC's 2025 Responsible AI Survey of 310 US business leaders found that maturing organizations are moving from shared committees to clear lines of accountability — with 56% of executives reporting that first-line teams (IT, engineering, data, and AI) now lead Responsible AI efforts. The survey recommends applying a three-lines-of-defense model to align builders, reviewers, and assurers — a model that has direct implications for how enterprises structure their vendor governance obligations.

Vendor governance, in other words, becomes an ongoing operational responsibility rather than a one-time procurement exercise. Procurement teams that have traditionally managed contract terms at the point of purchase now need to establish continuous oversight mechanisms that evolve alongside vendor platforms.

Another important shift is the movement of AI ethics from abstract discussion into operational governance. For many organizations, ethics initially appeared as a conceptual concern — centered around fairness, transparency, and responsible AI principles. But once AI systems begin influencing actual learning experiences, coaching environments, assessments, or employee development pathways, ethical concerns become operational realities.

This transition is accelerating under regulatory pressure. The EU AI Act, officially in force since August 2024, classifies AI systems used in employment decisions — including performance evaluation, work allocation, and learning progress assessment — as high-risk. Full compliance obligations for high-risk AI systems take effect from August 2026, requiring mandatory risk assessments, bias testing, human oversight mechanisms, and transparency disclosures. Non-compliance carries penalties of up to €35 million or 7% of global annual turnover.

Even outside the EU, the implications are significant. The "Brussels Effect" — whereby EU regulations effectively become global standards as multinational organizations apply them enterprise-wide — means that governance decisions made around EU AI Act compliance will shape AI use in learning environments globally, regardless of jurisdiction.

Emerging Governance Responsibilities Around AI in Learning

As governance complexity increases, many organizations are beginning to establish more formal oversight mechanisms — AI governance councils, responsible AI committees, cross-functional steering groups, and enterprise AI review boards. What is particularly significant about these structures is not their form but what they reveal organizationally: a growing recognition that no single department can fully govern AI alone.

According to Gartner data cited in a 2025 enterprise governance guide, over 60% of enterprises will require formal AI governance frameworks by 2026 to meet rising security, risk, and compliance demands. The 2025 IAPP Governance Survey found that only 28% of organizations have formally defined oversight roles for AI governance, highlighting persistent uncertainty about who owns responsibility for compliance, ethics, and model accountability — with most companies still distributing governance tasks across compliance, IT, and legal teams without a unified structure.

Looking at 2026 L&D trends, research from iVentiv shows the focus is already shifting from what AI can do to how it should be used responsibly — with clear governance, human oversight, and alignment to organizational values emerging as the critical differentiators. Learning culture has risen sharply as a CLO priority, from 16% in 2022 to 48% in 2025, and AI governance is central to this shift.

"If 2025 was about accelerating experiments, 2026 looks like a year where CLOs are asked to industrialise what works and redesign what doesn't — not just to keep up, but to ensure the organisation is still building the capabilities and culture that make performance sustainable."

— iVentiv Pulse Report, 2026 L&D Trends

At a deeper level, many of the governance tensions surrounding AI emerge because organizations are still attempting to apply governance models designed for stable software systems to technologies that are adaptive, generative, and continuously evolving.

Traditional governance models assumed predictable functionality, stable outputs, slower technology evolution, and clearer ownership boundaries. AI systems challenge all of these assumptions simultaneously. McKinsey's 2026 AI Trust Maturity Survey — covering approximately 500 organizations — found the average Responsible AI maturity score at just 2.3 out of 4, with only about one-third of organizations reporting maturity levels of three or higher in strategy, governance, and agentic AI governance. The survey concludes that while technical capabilities are advancing, organizational alignment and oversight structures are struggling to keep pace.

Common governance challenges identified by Responsible AI Labs include fragmented ownership (43% of organizations), unclear accountability (39%), lack of technical expertise (52%), difficulty measuring AI risks (47%), and cross-functional coordination challenges (41%). These are structural problems, not capability gaps — and they require structural solutions.

If AI governance increasingly cuts across organizational boundaries, then ownership itself may need to be redefined. Rather than asking "Which department owns AI?", organizations may need to ask "How do multiple functions govern AI together responsibly?"

This requires a shift from isolated ownership models toward coordinated governance models. Based on the research landscape and the patterns emerging across enterprise learning organizations, several governance shifts are becoming increasingly important:

⇄ Shared governance rather than siloed ownership

AI systems increasingly require collaborative accountability across functions. No single department can own systems that simultaneously touch infrastructure, learning, data, and organizational risk.

↗ Clear escalation and accountability pathways

Distributed governance still requires clearly defined responsibility structures. Shared governance without escalation paths creates ambiguity, not flexibility.

↻ Continuous vendor governance processes

Vendor oversight must evolve alongside changing AI systems. Procurement is no longer a one-time event — it is an ongoing governance discipline.

⚖ Operational ethics mechanisms

Responsible AI principles need practical implementation inside workflows — not policy documents. This includes bias monitoring, human oversight design, and transparent learner communication.

◎ Cross-functional review structures

Governance increasingly depends on coordinated expertise rather than isolated approvals. AI governance councils that bring L&D, IT, legal, security, HR, and procurement together are becoming standard practice among maturing organizations.

These are not small adjustments. They reflect a broader transition from tool governance to ecosystem governance — and from ownership to alignment.

One of the most important things AI is revealing inside workplace learning is that governance is no longer a peripheral issue attached to technology implementation. It is becoming central to whether AI adoption succeeds, scales, or stalls.

The challenge organizations face is not simply deciding which function owns AI. In many cases, no single function can fully own systems that simultaneously influence infrastructure, learning experience, workforce capability, analytics, operational risk, and organizational trust. McKinsey has found that fewer than 25% of companies have board-approved, structured AI policies — a governance gap that becomes harder to sustain as AI moves deeper into core workflows.

This means the real governance problem is not ownership alone. It is alignment — between L&D and IT, experimentation and accountability, innovation and oversight, vendor capability and enterprise control, operational flexibility and governance consistency.

The organizations that navigate this transition most effectively may not necessarily be the ones with the most advanced AI systems. They may be the ones that learn how to coordinate governance across functions in ways that are flexible enough to support innovation while structured enough to maintain trust.

Because ultimately, AI governance in learning is not simply a technology issue. It is an organizational design issue.

—RK Prasad (@RKPrasad)