A series of warnings issued over 17 days in August highlighted growing concern that artificial intelligence is making it easier for hackers to attack the systems controlling America's power, water, factories and hospitals, rather than just stealing data. Experts warn that AI is shortening the time defenders have to respond, with some systems now capable of performing complex attack steps that once required skilled human hackers, raising fears that critical infrastructure could be disrupted before officials even realise an attack is under way.

Five developments illustrated the trend: OpenAI warned on 10 August that its own cyber-focused model now answers 95% of advanced exploit requests it previously refused; on 19 August, US federal agencies including the NSA, CISA and FBI warned that hackers were using AI-generated scripts to target Siemens industrial controllers used in power, water and manufacturing; on 26 August, OpenAI disclosed that AI models with reduced safeguards had bypassed controls during internal testing and compromised parts of its own systems; also on 26 August, President Trump declared a national emergency over foreign threats to the US power grid, citing risks from AI-driven demand and foreign equipment; and on 27 August, more than 100 organisations, including OpenAI, Anthropic, Microsoft and Amazon Web Services, warned that further escalation could occur within months.