The rapid integration of artificial intelligence into corporate infrastructure has created a paradox of efficiency: systems are being deployed faster than they can be understood or controlled. According to a new report from ISACA, the global professional association for IT governance, a significant majority of organizations are operating without a viable "kill switch." Nearly 60 percent of digital trust professionals surveyed admitted they do not know how quickly their organization could halt an AI system during a security incident.
This lack of agency is more than a technical hurdle; it is a structural vulnerability. Only 21 percent of respondents indicated they could meaningfully intervene within 30 minutes of a system failure. In the high-stakes environment of automated decision-making and critical workflows, such a delay allows corrupted or malfunctioning models to operate unchecked, potentially leading to cascading operational failures or irreversible data damage.
The problem stems from a missing layer of governance. As Ali Sarrafi, CEO of the autonomous enterprise platform Kovant, notes, many businesses have embedded AI into their core operations without the necessary tools to supervise or audit their actions. When a system becomes a "black box," the ability to explain its behavior to regulators or even identify who is accountable for its errors vanishes. For many enterprises, the promise of automation has come at the cost of basic institutional control.
With reporting from AI News.
Source · AI News
