AI & Enterprise
Silent failures at scale could become AI risks that shake companies
As AI systems take on tasks such as transaction approvals, coding, customer support and data transfers, companies are seeing more cases where systems behave in unexpected ways. Experts say failures can occur quietly, making problems slow to detect, and that complex systems are hard to understand or quickly halt. They add that risks cannot be eliminated and call for stronger operational controls, supervision mechanisms and kill switches able to stop networks, not just apps.
