• European Commission President Ursula von der Leyen urged an EU push on self‑driving tech with AI pilots across European cities, positioning mobility as a priority for competitiveness and safety. Reuters

  • Von der Leyen also said the EU will table a unified “28th regime” for start‑ups in 2026 to simplify scaling across borders. Reuters

  • In the United States, Axios flagged a Commerce Department “50% rule” that could expand export‑control licensing duties to foreign subsidiaries majority‑owned by restricted entities, with notable implications for AI supply chains. Axios

  • A U.S. policy proposal for independent AI safety certification panels advanced in several states, offering limited liability for firms that meet higher safety standards. Axios

Regulation

  • The European Commission opened consultation on draft guidance and a template for reporting serious AI incidents under the AI Act, open until 7 November 2025. Compliance teams should align reporting playbooks now. Digital Strategy+1

  • The UK Ministry of Defence published its Responsible AI Senior Officers’ 2025 report, outlining governance and assurance practices for defence AI programmes. GOV.UK

Cases

  • OpenAI moved to dismiss xAI’s trade‑secret suit in California, calling it baseless, a fresh skirmish in the broader Musk–OpenAI litigation. Reuters

  • Studio plaintiffs filed new copyright actions targeting AI video‑generation products, alleging direct and secondary infringement, continuing the expansion of media IP litigation against GenAI tools. mckoolsmith.com

Events

  • AI Policy Summit at ETH Zurich, 3–4 October, convenes regulators and industry on governance and rights impacts. Center for Law & Economics+1

  • AI Liability Conference (TransJus and RAILS), Barcelona, 3 October, focuses on fault, causation and insurance for agentic systems. Universitat de Barcelona

  • Looking ahead: EDPB 109th Plenary on 7–8 October will touch governance coordination in the AI era. European Data Protection Board

Academia

  • SSRN highlighted Q3’s most‑downloaded “AI in Law” papers, a good snapshot of where legal scholarship is moving this autumn. SSRN Blog

  • New technical‑governance work on enterprise AI compliance frameworks appeared this week, useful for mapping policy to controls in large ERP environments. SSRN

Business

  • Export‑control expansion via the “50% rule” could force AI vendors to reassess partner and subsidiary structures, raising diligence costs for cross‑border model training and deployment. Axios

  • EU plans for a single start‑up rulebook and an AI‑first mobility push signal continued industrial policy aimed at keeping AI value chains in Europe. Reuters+1

Adoption of AI

  • Europe’s proposed city pilots for autonomous mobility indicate near‑term public‑sector adoption pathways that will test safety cases and data‑protection guardrails. Reuters

  • Defence governance: the UK MoD’s RAI report shows large organisations formalising stewardship roles and assurance processes for operational AI. GOV.UK

Takeaway

Industrial strategy and compliance are converging. Europe wants AI to power cars and scale‑ups, while the U.S. tightens export levers and the EU sharpens incident reporting. Governance teams should watch the export “50% rule,” stand up AI Act incident workflows, and track fast‑moving IP cases that continue to redefine acceptable training and outputs. 

Sources: Reuters: AI‑driven cars, Reuters: EU start‑up “28th regime”, Axios: Commerce “50% rule”, Axios: independent AI safety panels, European Commission consultation: serious incident reporting, UK MoD Responsible AI report 2025, Reuters: OpenAI move to dismiss xAI suit, McKool Smith: media copyright cases update, ETH Zurich AI Policy Summit, TransJus–RAILS AI Liability Conference, SSRN blog, SSRN: AI governance compliance framework