UpTrajectory Review

AI Accountability Gap Challenges IT Leaders

octopus tangled in computer cables
Photo: https://kaboompics.com/ / Pexels

UpTrajectory Review

The IBM Institute for Business Value surveyed CIOs and CTOs about who owns the AI systems now multiplying inside their companies, and the answer is a mess. Two-thirds of IT leaders say they remain accountable for AI they do not fully control, because employees and other departments keep building agents without IT's involvement. Seven in ten report that their organizations deploy technology faster than IT can track. CIOs expect agent counts to grow 38% within a year, yet only one in ten feels ready for that scale. Matt Lyteson, IBM's CIO of technology transformation, frames the problem as a governance gap: enterprises have opened agent development to everyone, not just the IT department, and the control mechanisms have not caught up.

For a small-business operator, this story will sound less like enterprise news and more like Tuesday. A five-person shop faces the same trap when the marketing hire connects a chatbot to the content calendar or the bookkeeper pastes projections into an LLM. Nobody asked IT, because there is no IT. The owner still carries the liability if the tool leaks customer data, invents a forecast, or violates a client contract. The IBM numbers simply confirm that even companies with dedicated technology staff have lost the thread. If two-thirds of surveyed CIOs cannot map the agents running inside their walls, a business owner with a shared password and a credit card has no chance without deliberate effort.

The genuinely useful insight here belongs to Ben Schein of Domo: the pace problem is not recklessness. Employees act reasonably. They solve real workflow pain with tools IT has not approved or even noticed. That reframing matters, because the usual response, a blanket ban on unauthorized AI, treats a governance failure as a discipline problem and usually fails. We agree with Schein that the gap lies in governance models, not in employee intent. We remain skeptical, however, of any suggestion that better dashboards alone will close it. Inventory tools tell you what exists; they do not tell you whether the agent's output is accurate, secure, or worth the subscription fee.

The downstream costs split unevenly. Large enterprises absorb the risk through compliance teams, insurance, and fired middle managers. Small businesses absorb it directly: a hallucinated quote sent to a customer, a data-privacy complaint, a workflow that depends on a vendor's API and breaks overnight. The 38% growth projection also signals budget creep. Agents rarely replace existing software; they layer on top, which means more licenses, more integrations, and more surface area for failure. Meanwhile the employees who built the agents move on, leaving the owner holding an undocumented automation nobody understands.

Watch whether your own operation can produce, within one hour, a list of every AI tool touching customer data, money movement, or external communication. If you cannot, start there before you buy one more agent. Assign one person, even if that person is you, to review each tool for what data it touches and who approved it. Set a simple rule: no new AI agent without a named owner and a written note on what it does. The CIOs in the survey wish they had done that six months ago. A small business can still do it this afternoon.

Takeaway: List every AI agent touching your customer data or money today, assign each a named owner, and require written approval before anyone adds another.

Excerpt from the original — CIO Magazine

IT leaders are getting a sneak preview of governance in the agentic era, and it’s shaping up to be a horror show.

Two-thirds of CIOs and CTO surveyed by the IBM Institute for Business Value say they’re accountable for AI systems they don’t fully control as employees and other business units spin up new agents.

In addition, 70% of the IT leaders surveyed say their organizations are deploying tech systems faster than their IT teams can track. Moreover, CIOs expect a 38% increase in the number of AI agents deployed at their organizations by next year, with just one in 10 IT leaders saying they’re prepared for the anticipated scale of agent deployment.

The study reinforces growing concerns about the state of AI governance, with many IT organizations challenged to track output, security, and value as employees spin up new agents without IT’s input, says Matt Lyteson, CIO of …