Building government resilience in an era of AI-driven cyberattacks

With reports showing an 87% surge in artificial intelligence (AI)-driven cyberattacks in 2025 and one in six breaches involving methods where AI was used to automate and amplify attacks, 2026 will be a make‑or‑break year for governments. Organizations must generationally evolve their defenses and plan for destructive AI-based attacks, particularly by foreign adversaries. 

Understanding the AI threat landscape

AI-driven attacks are extremely difficult to defend against due to their velocity, frequency and unpredictability. Attackers can shape outcomes, undermine trust and prepare the cyber battlespace long before any overt disruption occurs. These attackers are quickly and quietly compromising the datasets that support critical government decision-making, operations and infrastructure.

State‑linked groups like Salt Typhoon exemplify this shift by focusing on long‑term, covert access rather than quick, noisy wins. During these hacks, as we saw with the Anthropic hack in September 2025, AI agents blend in with normal ...