G7 AI Summit: towards an international governance framework — what will change

Cardan-AI decoded the conclusions of the G7 AI Summit of May 2026, held under the Canadian presidency. The seven countries adopted a common framework built around three principles: algorithmic transparency for systems with societal impact, developer accountability for foreseeable uses of their models, and users' rights to a comprehensible explanation of automated decisions affecting them. This framework is non-binding at this stage, but it foreshadows regulatory developments in the coming years.

Cardan-AI's analysis of the implications for multinationals: organisations operating across multiple G7 countries will be the first to feel regulatory convergence. Transparency and explainability requirements will progressively align between the European AI Act, the Canadian framework (AIDA), and US executive orders. Anticipating this convergence by documenting AI system design choices now is the best long-term compliance strategy.

The most operational point in the G7 framework concerns 'rights to explanation'. By 2027–2028, any user affected by a significant automated decision (credit, employment, insurance, service access) should be able to obtain a comprehensible explanation. This directly impacts scoring, recommendation, and HR decision-support systems that many companies have deployed in recent years.

Cardan-AI recommends conducting an immediate audit of existing automated decision systems, identifying those that will be subject to explainability requirements, and planning their adaptation. Implementing XAI (Explainable AI) techniques is technically mature — the real challenge is organisational and documentary.