OECD launches streamlined Hiroshima AI Process Reporting Framework to help small and medium-sized enterprises participate

The OECD has launched version 2.0 of the voluntary reporting framework of the Hiroshima AI Process. The launch took place at a Tech7 event held alongside the G7 Digital and Tech Ministerial Meeting in Paris, under the French G7 Presidency. The announcement calls it the only global framework through which companies describe their work toward safe, secure and trustworthy AI.

The new version is described as simpler and meant to make participation easier for actors across the AI ecosystem, with particular attention to small and medium enterprises. The revisions move the framework's reach beyond large developers of AI models to the wider group of AI developers and deployers. Over 50 companies have already committed to completing it, and it builds on a first year described as successful.

The framework helps monitor how the Hiroshima Process code of conduct for organisations building advanced AI systems is applied. That code is a central part of the Hiroshima AI Process, which began under Japan's 2023 G7 Presidency. Organisations that contribute give comparable information about their AI practices, which the announcement says fosters trust and accountability and helps spread good practices. By matching the framework to several risk management systems, the OECD says it aims to encourage interoperability and consistency of AI systems worldwide.

OECD Deputy Secretary-General Yasushi Masaki is quoted saying that wider sharing of practices will support AI adoption across industries and firm sizes, and that the improved interface will especially help small and medium enterprises. Submissions are accepted on a rolling basis, and organisations are invited to update their reports each year.