Hiroshima AI Reporting Framework

The Hiroshima AI Reporting Framework is open to organisations across the AI value chain, including developers, deployers, and providers of advanced AI systems. Participants are invited to report on their practices for managing risks and advancing trustworthy AI. Participation strengthens transparency, contributes to a shared understanding of organisational practices, and enables organisations to demonstrate their approaches to trustworthy AI.

OverviewAboutFAQs
FeedbackHAIP Brand
Submit your Report
Submit your Report

Frequently Asked Questions

What is the objective of the reporting framework?

The Reporting Framework promotes transparency and accountability among organisations across the AI value chain. This framework operates through the HAIP International Code of Conduct, which reflects the principles outlined in the Hiroshima Artificial Intelligence Process (HAIP) and builds on the OECD AI Principles.

What's new in version 2.0?

Refined through a pilot involving organisations from seven countries, including model developers, application developers, and deployers across the AI value chain, alongside input from individual experts, the new version:

  • Enhances access for organisations of all sizes, especially small and medium enterprises, with a simpler and more structured reporting experience.
  • Reflects organisations' roles across the AI system lifecycle, distinguishing between model developers, application developers, and deployers so each organisation can focus on the questions most relevant to its role.
  • Connects to the OECD.AI Catalogue of Tools and Metrics for Trustworthy AI, so organisations can reference tools they already use and contribute new ones, supporting the exchange of good practices.
  • Keeps pace with new technology, including agentic AI and other emerging capabilities.

By aligning with the Hiroshima Code of Conduct, as well as other widely used risk management approaches and voluntary commitments, Version 2.0 also supports interoperability and consistency across the global AI landscape.

Who is eligible to participate in the reporting framework?

Organisations must be based in a jurisdiction that meets at least one of the following conditions:

What are the submission requirements?

  • Organisations must attest that all information provided is accurate and true to the best of their knowledge at the time of submission.
  • Organisations may be requested to provide complementary submissions between reporting periods.
  • Organisations must answer all questions in the reporting framework.

How will submissions be processed?

  • The OECD will make submissions publicly available in full on a dedicated webpage on oecd.ai.
  • The OECD may request additional non-public supporting information.
  • The OECD Secretariat will verify that:
    • The organisation submitting the report is based in an eligible jurisdiction.
    • All questions are answered.
    • Supporting materials (e.g., links) are accessible.
  • The Secretariat will not assess or verify the substance of submissions.

What does it mean to be listed under the HAIP Brand?

Organisations that commit to the HAIP International Code of Conduct and complete the reporting framework are recognised as participants in the HAIP Reporting Framework, with the HAIP Brand displayed on their published report. Reports are published on the OECD.AI Policy Observatory. To continue being recognised as active participants, organisations are expected to update or resubmit their report annually (i.e., at least every 12 months).

What should organisations know about the HAIP Brand listing?

Listing under the HAIP brand does not constitute:

  • An endorsement of the organisation's practices or the AI systems it develops or uses.
  • A certification of compliance with the HAIP International Code of Conduct for Organisations Developing Advanced AI Systems.

Under what circumstances might an organisation be delisted?

Organisations may be delisted from the HAIP Brand webpage under any one of the following circumstances:

  • It has been discovered that the submissions did not come from an organisation based in an eligible jurisdiction.
  • The organisation fails to meet the reporting frequency requirements (i.e. has not completed the reporting framework in over a year).
  • The organisation fails to respond to requests for complementary submissions within the required timeframe.
  • It is found that the organisation acted in a manner that is not in line with its submission under the HAIP International Code of Conduct Reporting Framework, or it provided information that is not accurate and true to the best of its knowledge.