86% of CIOs say AI risk outpaces value.

The same ungoverned agents are being sold to governments with fewer safeguards.

Unlike past software, AI agents don’t leave predictable audit trails and their permissions often stay invisible to existing identity systems.

Gartner urges firms to build “guardian agents” with power to kill rogue AI processes.

How each outlet framed it

drawn from 30+ reports worldwide

TheRegister.com leans critical
indicts AI vendors' immaturity—frequent model updates break applications, six-month lifespan, no legacy support despite enterprise dependency

Sources: TheRegister.com