In COTS AI deployments, which factor most directly affects accountability?

• A. Unclear Definition of Responsibilities for Maintenance and Controls
• B. Clear SLAs
• C. No Documentation
• D. Minimal Governance

Answer: (A)

Accountability in COTS AI deployments hinges on who owns the ongoing maintenance, monitoring, and controls for the system. When responsibilities for maintenance and controls are not clearly defined, there’s no single owner to answer for updates, model changes, data handling, monitoring results, or incident response. This ambiguity makes it easy for issues to be ignored or for blame to be shifted, and it undermines the ability to enforce policies, audits, and regulatory requirements. Even if SLAs exist or there is documentation and governance elsewhere, without a clearly assigned owner for maintenance and controls, accountability cannot be effectively established or enforced. Clearly defined responsibilities ensure there is a accountable party responsible for keeping the AI solution secure, compliant, and reliable, and for addressing problems when they arise.