What happened
The OECD released Version 2.0 of the Hiroshima AI Process (HAIP) Reporting Framework on 28 May 2026, at an event hosted by Tech7 on the margins of the G7 Digital and Tech Ministerial Meeting in Paris under the French G7 Presidency. The update is substantive: the original v1 framework applied a single uniform questionnaire to all AI developers, whereas v2.0 introduces role-differentiated reporting that distinguishes model developers, application developers, and deployers, so each organisation answers only the questions relevant to its position in the AI value chain. Version 2.0 also adds coverage of agentic AI and other emerging capabilities, links directly to the OECD.AI Catalogue of Tools and Metrics for Trustworthy AI so organisations can cite existing tools, and streamlines the experience for small and medium enterprises — a deliberate expansion of the participating population. The revision was refined through a pilot involving organisations from seven countries across the AI value chain. More than 50 organisations — including Google, OpenAI, Fujitsu, Salesforce, NEC, and Rakuten — have already pledged to submit reports under the new version, with a 1 September 2026 deadline to be included in the first consolidated v2.0 publication. As the only international framework for organisations to report on their efforts to promote trustworthy AI, Version 2.0 directly implements the Hiroshima Process International Code of Conduct for Organisations Developing Advanced AI Systems agreed at Japan's 2023 G7 Presidency.
Why it matters
Any AI lab, enterprise deployer, or government procurement team operating across G7 jurisdictions now faces a reporting standard that is actively enrolling major vendors; compliance with or deviation from HAIP v2.0 will increasingly appear in procurement, regulatory, and board-level due diligence conversations. The role-differentiated structure means deployer organisations — not just model developers — have a clearly scoped reporting obligation for the first time.
Action needed
Policy and compliance leads should map their organisation's position in the AI value chain (model developer, application developer, or deployer) against the v2.0 questionnaire on oecd.ai/transparency and assess whether a voluntary submission before the 1 September 2026 cohort deadline is strategically appropriate.