Overview
- G7 finance ministers agreed to run simulations to probe how authorities would respond to cyberattacks that use AI tools, a step designed to test protocols for banks, central banks and finance ministries.
- Google DeepMind executive Demis Hassabis and other tech leaders warned this week that autonomous agents can behave unpredictably and that attacks using AI could grow more damaging as systems gain capability.
- Companies are deploying AI fast inside operations, often without full controls, producing widespread 'shadow AI' where employees use external models and chatbots in ways that risk exposing sensitive data.
- Banks and large firms are moving from pilots to production with highly autonomous agents and ‘factories of agents’ supervised by small human teams, a shift that boosts efficiency but concentrates systemic risk in a few providers and data centers.
- Trade and investment are reshaping around AI: demand for chips, data‑center equipment and networking drove a large share of global trade growth in early 2026, which increases the economic leverage of the actors that control that infrastructure and raises the stakes for coordinated governance and legal accountability.