Why Leaders Dodge The Hot Potato

When organizations debate who should own AI, the conversation quickly fragments into role narratives rather than a governance blueprint.

CIOs point to platforms and operational scale. CTOs emphasize product integration and experimentation. CFOs measure risk to the balance sheet. COOs view AI as an operations transformation. Compliance or legal teams see regulatory exposure.

Each perspective is defensible because AI touches data, engineering, operations, finance and people simultaneously. However, that very overlap makes it easy to diffuse responsibility and hard to write a single owner into a clear accountability structure. As a result, this joint “ownership by committee” becomes shared neglect, and AI programs either stall in pilots or proliferate in an unmanaged fashion across multiple platforms.

The technical reason for the hot potato is vendor and platform fragmentation. Enterprises routinely run multiple “primary” AI platforms with different controls, SLAs and transparency, and few organizations have someone with the mandate and the teeth to enforce consistent standards across those vendors.

The absence of a central owner who can demand vendor transparency and ownership of outcomes creates both operational risk and vendor lock-in, which no single executive wants to be blamed for when it surfaces in an audit or a PR crisis.

How Boards And CEOs Can Reclaim Their Roles

My work on the boards of companies has shown me how it can work differently. For many startups, AI is essential to survival, making ownership more than technical stewardship. It’s about the authority to arbitrate competing priorities across the dimensions of strategy, risk, talent and investment.

That is why, for me, the CEO must lead the design of the AI strategy as part of the wider business strategy, and the board must ultimately approve it. This strategic focus shows that AI is not merely infrastructure. It’s a development that reshapes products, customer relationships and workforce design.

Without a C-suite sponsor empowered to align incentives, tie outcomes to compensation and resolve trade-offs quarterly, AI becomes a mosaic of local optimizations rather than a source of enterprise differentiation.

Boards also need better, more frequent reporting on AI risk and impact. High-level slide decks won’t do. Directors must receive concrete metrics on model performance, data lineage, vendor dependencies and incident response readiness. When boards treat AI as a governance afterthought, they cede the narrative to technical teams and vendors. Thus, they increase the risk of surprises landing on the CEO’s desk with little context or a remediation plan.

Operational Design: Built So Nobody Gets Burned

Practical ownership does not mean a single monolithic AI office making decisions. Instead, companies that succeed separate accountability from delivery.

Accountability, the authority to set strategy, approve budgets and arbitrate trade-offs, should sit with a senior executive reporting to the CEO, with written delegation and board oversight.

Delivery is distributed among a coordinated set of stewards. The CIO for platform and security, CDO for data stewardship, CTO for product integration and business leaders for outcomes. When each party’s responsibilities are explicit, reviewed regularly and tied to business outcomes, joint ownership becomes workable rather than a way to dodge risk.

Equally important is creating a pragmatic governance fabric. Companies need lightweight but enforceable standards, vendor playbooks, model registries and incident protocols. They must be simple enough to operationalize but also be rigorous enough to provide auditability.​

Failure often doesn’t stem from technical immaturity but operational diffusion. Without a repeatable operating model, you can’t safely convert your pilots into production at scale.

The Cultural Consequences Of Passing The Hot Potato

Organizations that treat AI as someone else’s problem also create cultural hazards. Engineering and data teams become frustrated by ambiguous priorities, while risk and legal teams feel sidelined until a crisis pops up. Meanwhile, business units either over-deploy unchecked models or under-utilize safe, value-creating capabilities.

In all cases, the enterprise loses. Either through avoidable compliance incidents or through missed economic opportunities because leaders never had the conviction to standardize and scale successful pilots.

In my experience, making ownership explicit, aligning incentives and treating governance as a competitive advantage rather than a compliance chore helps alleviate all facets of this problem. Not only do teams gain clarity, but individual contributors get a clear mandate to take charge, control the technology and grow. All of which ultimately helps the company stay in control of its AI projects.​

Grab The Spud And Start Taking Responsibility

The AI hot potato will only stop being tossed around when organizations accept that ownership is less about technology and more about corporate purpose and accountability.

When CEOs and boards treat AI as a strategic asset, and when governance frameworks allocate clear authority while empowering delivery teams, then AI stops being a political risk and becomes a managed capability that accelerates value.

Getting there requires decisive leadership, not just technical excellence. Until someone is willing to be accountable, AI strategy will remain the riskiest game of hot potato in the C-suite. And the leaders who grab it first and hold on are the ones who will define what successful AI actually looks like in practice.​

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?