Germany’s new AI safety institute will work with Canada, Digital Minister Karsten Wildberger announced Monday at a two-day government summit on AI in Berlin, as Chancellor Friedrich Merz warned that Europe is running out of time to catch up in AI. A government task force presented a report warning that the window for Germany to close the gap with the world’s AI leaders “is closing soon,” and set a goal of making Germany an “AI nation” by 2030. Merz said Germany was caught in a “storm of technological transformation” and that a single law can take so long to pass that three new generations of AI models are developed in the meantime. The institute, called AISI Deutschland, opened in Berlin in early September 2026 and is built around the Federal Office for Information Security and the Federal Network Agency; it is not a regulator but is meant to assess the risks of advanced AI models. No timeline for the cooperation with Canada was given. Merz said Germany had also helped launch a European Frontier AI Initiative and that cooperation with France and Canada would help Europe share knowledge and build alternatives to the dominant American AI models. He set out a two-track plan: in the short term, Germany will support AI built for industry; in the longer term, it will fund research into future systems, including AI-powered robots. The summit, which ends Tuesday, was expected to draw more than 1,000 guests, including digital ministers from across Europe and Canada and the heads of Mist