SynopsisUnchecked AI agents & runaway code sprawl are quietly driving a 2,500% explosion in software defects, along with security and governance risks, writes Himanshi Lohchab.The software bug should have taken minutes to fix. Instead, it took two days. The engineer debugging the failure had never seen the logic before. It had been generated by one AI model, modified by another and deployed automatically by an AI agent.As more companies rely on AI to write software, developers are increasingly finding themselves maintaining code that no human fully understands. The result is a new enterprise problem that technologyNow Playing