Back to blog

Why does no one own your AI risk until it blows up?

Thắng Đoàn
Thắng Đoàn

Your AI systems are already making decisions that can hurt customers and draw regulators. Yet at most organizations, nobody clearly owns the risk.

Accountability without authority

A CISO, or chief information security officer, is the executive responsible for security. TechTarget reports that new PwC research finds no consensus on who should own AI governance and risk, with responsibility scattered across security, technology, and split roles. Nemertes analyst Johna Till Johnson says a CISO's formal job is to set strategy and prevent harm, while their actual job is to be fired when something goes wrong. AI is now one of the biggest such somethings. Experts quoted by TechTarget say CISOs need authority over AI security, not just blame when it fails.

The chatbot liability trap

Unclear ownership is not just an org chart problem. American Banker reports that banks using AI chatbots for mortgage and loan support may risk breaking consumer-protection laws. The report's authors say corporations remain liable when chatbots give wrong answers, create endless doom loops, or fail to solve problems. Ownership of risk here means something concrete: bearing legal responsibility for outcomes. The bank pays even when a vendor's model produced the failure.

Regulators are closing in

Oversight is tightening before ownership questions are settled. PYMNTS reports that U.S. banking watchdogs, including the Federal Reserve among the top three regulators, are ramping up scrutiny of AI risks. Drivers include cyberattacks on banks and rogue AI agents threatening financial stability. Sources expect more focus on dependency on outside tech providers and more formal guidance in some jurisdictions. AFR reports bank chief risk officers are losing sleep over rogue AI agents causing real-time system issues or contagion. This is the ownership vacuum made urgent: regulators want a named accountable party before firms have one.

What this does not solve

Naming an owner does not make the risk measurable. None of the sources offer a tested framework for auditing AI agent behavior at scale. Ownership also will not resolve the tension between regulators, who want formal accountability, and firms that want to move fast. And clear internal ownership may not shield you from liability when a third-party model fails.

If your AI chatbot gave a customer wrong loan advice today, whose desk would the incident land on?
Do your risk owners have authority over AI deployments, or only accountability after failures?

Write down every AI system your firm runs in production and name the single owner of each. Any blank is your exposure.

Sources

TechTarget: Unclear AI risk ownership could leave CISOs in scapegoat role

American Banker: Banks face liability over use of AI chatbots

PYMNTS: Banking Regulators Tighten Oversight of Possible AI Risks

AFR: Bank chief risk officers are losing sleep over rogue AI agent threats

Share:

Recommended for you

Enjoyed this article?

Subscribe for new articles. No spam. Unsubscribe anytime.

By subscribing you agree to receive the newsletter. No spam, and you can unsubscribe anytime.