October digital policy roundup lists AI Act enforcement consultation and Japan’s new AI council

Digital Policy Alert, an independent repository tracking policy changes that affect the digital economy, published a roundup of October 2026 developments drawn from monitoring in more than 60 jurisdictions. Its artificial intelligence section opens with the European Commission's consultation on a draft Implementing Regulation laying down procedures for the Commission's AI Act enforcement, under Articles 75c and 75d. They would cover AI systems under the Commission's direct supervision, including those built on a general-purpose model by the same provider and those embedded in very large online platforms or in search engines. The draft also sets when limitation periods for fines start, and when they are interrupted or suspended.

Other European items include the harmonised AI Act standard on quality management systems becoming applicable, an EU-Canada joint statement on AI safety, security and governance, and Italy's Legislative Decree implementing the AI Act. That decree makes AgID the notifying authority, with the National Cybersecurity Agency acting as market surveillance body, sets fines up to 35 million euros or 7 percent of turnover, and requires a person to make final employment decisions. Finland, Poland and Belgium also advanced AI-related measures.

In Asia, Japan's Cabinet created the AI and Digital Reform Promotion Council. Its first meeting is to address revising trustworthy-AI guidelines, whether to amend the AI Promotion Act, who is liable for AI agents and physical AI, and oversight of advanced AI capabilities. China's ministries opened a consultation on adding 23 AI technologies to export controls, and South Korea's Fair Trade Commission consulted on guidelines against AI washing. The roundup also reports a renewed Canada-UK AI security cooperation arrangement, the UK Law Commission's product liability consultation and an ICO call for evidence on agentic AI.