Hiroshima AI Process

2nd in-person meeting of the Hiroshima AI Process Friends Group, Tokyo, 15–16 March 2026 — closing ceremony at which the Action Plan 2026 was announced. Source: Ministry of Foreign Affairs of Japan.
2nd in-person meeting of the Hiroshima AI Process Friends Group, Tokyo, 15–16 March 2026 — closing ceremony at which the Action Plan 2026 was announced. Source: Ministry of Foreign Affairs of Japan.

CAIDP is a member of the Hiroshima AI Process Friends Group Partners' Community

“CAIDP was recently designated as one of the first civil society members of the Hiroshima AI Process Friends Group.”

News

16 MARCH 2026 · TOKYO
HAIP Friends Group adopts the Action Plan 2026

At the closing ceremony of the 2nd in-person Friends Group meeting, the Hiroshima AI Process Friends Group Action Plan 2026 was announced. Parliamentary Vice-Minister for Foreign Affairs SHIMADA Tomoaki delivered the closing remarks.

29 OCTOBER 2025 · MIC, JAPAN
CAIDP joins the HAIP Friends Group Partners' Community

Japan’s Ministry of Internal Affairs and Communications announced that ten organizations, including the Center for AI and Digital Policy, had joined the Partners' Community.

From: Ministry of Internal Affairs and Communications of Japan, 29 Oct 2025
7 FEBRUARY 2025 · OECD
OECD HAIP Reporting Framework launches

The G7 Hiroshima AI Process Reporting Framework opened at transparency.oecd.ai. As of 4 May 2026, 25 organizations had published reports under the Framework, and a v2.0 pilot was open as of March 2026.

28 FEBRUARY 2025 · TOKYO
1st in-person Friends Group meeting; Partners' Community launched

Held at the Hotel New Otani. Parliamentary Vice-Minister for Foreign Affairs IKUINA Akiko delivered the closing remarks. The Partners' Community was announced on Day 2 with 16 founding organizations.

 

From: Ministry of Foreign Affairs of Japan, 28 Feb 202

66 Friends Group countries · 38 Partners' Community organizations

66FRIENDS GROUP COUNTRIES / REGIONS
38PARTNERS' COMMUNITY ORGANIZATIONS
25OECD HAIP REPORTERS
11CODE OF CONDUCT ACTIONS
As of 16 March 2026 for membership counts and 4 May 2026 for the OECD reporter count.

SELECTED PARTNERS' COMMUNITY MEMBERS

  • Center for AI and Digital Policy (CAIDP)
  • Adobe · Amazon · Anthropic
  • Box Japan · Business Software Alliance · Fujitsu
  • Google · Hitachi · IBM · Impact AI
  • Infosys · ISACA · JICA · KDDI
  • Microsoft · Mila · NEC · NTT
  • OECD · OpenAI · Oracle Japan · Palo Alto Networks
  • Preferred Networks · Rakuten · SaferAI · Salesforce · SoftBank
  • UNDP · United Nations University · World Bank · World Economic Forum

Full membership list (MIC, Japan)

FIRST WAVE OF PUBLIC HAIP REPORTS — 22 APRIL 2025

  • Anthropic · Google · Microsoft · OpenAI · Salesforce
  • NEC · NTT · Fujitsu · KDDI · SoftBank · Rakuten · Preferred Networks
  • ai21 · KYP.ai · Data Privacy and AI · TELUS · Fayston Preparatory School

All HAIP reports (OECD)



Background

The Hiroshima AI Process (HAIP) was launched at the G7 Hiroshima Summit in May 2023 under Japan’s G7 Presidency, with the objective of discussing the opportunities and risks of generative AI and promoting safe, secure, and trustworthy advanced AI.

On 30 October 2023, the G7 Leaders agreed the Hiroshima AI Process Comprehensive Policy Framework, comprising the International Guiding Principles for Organizations Developing Advanced AI Systems and the International Code of Conduct for Organizations Developing Advanced AI Systems.

Since 2024, HAIP has been operationalised through three structures: the Friends Group, launched by Japan in May 2024; the Partners' Community, launched in February 2025; and the OECD HAIP Reporting Framework, launched on 7 February 2025.

Successive G7 presidencies have reaffirmed and extended HAIP: Italy in 2024 at Trento, Apulia and Cernobbio; Canada in 2025 at Kananaskis and Montréal; and France in 2026.

Timeline

MAY 2023
G7 Hiroshima Summit launches HAIP
Leaders direct G7 ministers to develop an international framework for advanced AI systems.
30 OCTOBER 2023
G7 Leaders' Statement on the Hiroshima AI Process
Guiding Principles and Code of Conduct issued; Comprehensive Policy Framework agreed at the December 2023 G7 Digital & Tech Ministers' Meeting.
15 MARCH 2024
Trento Ministerial Declaration
Annex 3 dedicates the G7 to advancing HAIP outcomes; the OECD is tasked with developing monitoring tools for the Code of Conduct.
MAY 2024
HAIP Friends Group launched by Japan
A voluntary framework of countries supporting HAIP.
13–15 JUNE 2024
Apulia G7 Leaders' Communiqué
Leaders reaffirm commitment to the Reporting Framework with OECD support.
9 JULY – 6 SEPTEMBER 2024
OECD pilot of the Reporting Framework
20 organizations across 10 countries participate in the pilot.
15 OCTOBER 2024
Cernobbio G7 Tech and Digital Ministerial Statement
Pilot results reviewed; G7 commits to advance the Framework by year-end.
7 FEBRUARY 2025
OECD HAIP Reporting Framework launches
Live transparency portal at transparency.oecd.ai goes online.
27–28 FEBRUARY 2025
1st in-person HAIP Friends Group Meeting
Partners' Community launched on Day 2 with 16 founding organizations.
22 APRIL 2025
First wave of public HAIP reports
17 organizations including Anthropic, OpenAI, Google, Microsoft, Salesforce, NEC, and Preferred Networks publish on transparency.oecd.ai.
2 JUNE 2025
CAIDP files G7 Summit statement
CAIDP states that it was recently designated as one of the first civil society members of the Hiroshima AI Process Friends Group.
17 JUNE 2025
G7 Leaders' Statement on AI for Prosperity
Launches the G7 AI Adoption Roadmap; its trust-building pillar is grounded in HAIP.
29 OCTOBER 2025
MIC press release confirms CAIDP in Partners' Community
Announced alongside UNDP, WEF, BSA, Anthropic, Adobe, Oracle Japan, Box Japan, Impact AI, and Palo Alto Networks.
9 DECEMBER 2025
G7 IDT Ministerial Declaration
Canadian Presidency confirms continued uptake of the Code of Conduct Reporting Framework.
15–16 MARCH 2026
2nd in-person HAIP Friends Group Meeting
Action Plan 2026 announced. 66 countries and 38 Partners' Community organizations are listed.

 

Core Documents (2023, Japanese Presidency)

G7 Leaders' Statement on the Hiroshima AI Process — 30 October 2023

SOURCE: MINISTRY OF FOREIGN AFFAIRS OF JAPAN

The G7 Leaders endorse the Comprehensive Policy Framework, attaching the International Guiding Principles and the International Code of Conduct.

MOFA — anchor page

Hiroshima Process International Guiding Principles for Organizations Developing Advanced AI Systems

SOURCE: MINISTRY OF FOREIGN AFFAIRS OF JAPAN

11-point set of voluntary principles for organizations developing advanced AI systems. Companion to the Code of Conduct.

PDF (337 KB)

Hiroshima Process International Code of Conduct for Organizations Developing Advanced AI Systems

SOURCE: MINISTRY OF FOREIGN AFFAIRS OF JAPAN

11 voluntary actions across the AI lifecycle: risk identification, post-deployment vulnerability monitoring, public reporting, information sharing, AI governance and risk management, security controls, content provenance, safety research, AI for global challenges, technical standards, and data input safeguards.

PDF (274 KB) · Full text on MIC Documents of Achievement

G7 Italian Presidency (2024)

Trento Ministerial Declaration on Industry, Technology and Digital — 15 March 2024

SOURCE: G7 ITALIAN PRESIDENCY

Annex 3 is dedicated to HAIP and tasks the OECD with developing monitoring tools for the Code of Conduct.

Identify, develop, and introduce appropriate tools and mechanisms for monitoring the application of the Code of Conduct by organisations that will commit to these outcomes on a voluntary basis… — Trento Declaration, Annex 3

G7 Research Group mirror

Apulia G7 Leaders' Communiqué — 13–15 June 2024

SOURCE: G7 ITALIAN PRESIDENCY

G7 Leaders reaffirmed their commitment to developing the HAIP Reporting Framework in cooperation with the OECD. The OECD subsequently ran a pilot from 9 July to 6 September 2024 with 20 organizations across 10 countries.

Apulia Communiqué PDF

Cernobbio G7 Tech and Digital Ministerial Statement — 15 October 2024

SOURCE: G7 ITALIAN PRESIDENCY (REPUBLISHED BY GOV.UK)

G7 Digital and Tech Ministers reviewed the OECD pilot results and committed to advancing the Reporting Framework by year-end.

We discussed the findings from the pilot phase of the reporting framework for the Hiroshima Process International Code of Conduct for organisations developing advanced AI systems, conducted by the OECD from 19 July to 6 September 2024. — Cernobbio Statement, 15 October 2024

GOV.UK republication

G7 Canadian Presidency (2025)

G7 Leaders' Statement on AI for Prosperity — Kananaskis, 17 June 2025

SOURCE: OFFICE OF THE PRIME MINISTER OF CANADA

G7 Leaders launched the G7 AI Adoption Roadmap. Its trust-building pillar is explicitly grounded in HAIP.

We plan to build on progress achieved under the Japanese and Italian presidencies and leverage the outcomes of the Hiroshima AI Process (HAIP) to foster trust. — G7 AI Adoption Roadmap annex, Kananaskis, 17 June 2025

G7 Research Group mirror

G7 2025 Industry, Digital and Technology Ministerial Declaration — Montréal, 9 December 2025

SOURCE: G7 CANADIAN PRESIDENCY

G7 IDT Ministers asked the OECD to make the published HAIP reports easier for users and stakeholders to understand.

We recognize the role of the Code of Conduct Reporting Framework developed under Italy's 2024 Presidency for facilitating user trust by reinforcing accountability and providing transparency of advanced AI systems through voluntary actions by AI providers, as well as the interest in its continued uptake. — G7 2025 IDT Declaration, Montréal

G7/G20 Documents Database · Canadian primary

Implementing the G7 AI Adoption Roadmap — Montréal, 9 December 2025

SOURCE: G7 CANADIAN PRESIDENCY

Country-by-country implementation snapshot of the Roadmap. The “Unlock AI Opportunity Through Trust-building” section is grounded in HAIP.

G7/G20 Documents Database

Friends Group In-Person Meetings

2ND IN-PERSON FRIENDS GROUP MEETING — TOKYO, 15–16 MARCH 2026

Held at the Hotel New Otani. The HAIP Friends Group Action Plan 2026 was announced at the closing ceremony. Closing remarks by Parliamentary Vice-Minister for Foreign Affairs SHIMADA Tomoaki.

Parliamentary Vice-Minister for Foreign Affairs SHIMADA Tomoaki at the closing ceremony of the 2nd in-person Hiroshima AI Process Friends Group Meeting, Tokyo, 16 March 2026.
Parliamentary Vice-Minister for Foreign Affairs SHIMADA Tomoaki at the closing ceremony of the 2nd in-person Hiroshima AI Process Friends Group Meeting, Tokyo, 16 March 2026. Source: Ministry of Foreign Affairs of Japan.
Scenes from the closing ceremony of the 2nd in-person Hiroshima AI Process Friends Group Meeting, Tokyo, 16 March 2026.
Scenes from the closing ceremony of the 2nd in-person Hiroshima AI Process Friends Group Meeting, Tokyo, 16 March 2026. The Action Plan 2026 was announced at this meeting. Source: Ministry of Foreign Affairs of Japan.

MOFA press release, 16 March 2026 · HAIP Friends Group Action Plan 2026 (PDF)

1ST IN-PERSON FRIENDS GROUP MEETING — TOKYO, 27–28 FEBRUARY 2025

Held at the Hotel New Otani. Day 1 included a video message from Prime Minister Ishiba, a keynote from Prof. Matsuo of the University of Tokyo, and site visits to NEC and NICT. Day 2 announced the launch of the Partners' Community. Closing remarks by Parliamentary Vice-Minister for Foreign Affairs IKUINA Akiko.

Parliamentary Vice-Minister for Foreign Affairs IKUINA Akiko at the closing ceremony of the 1st in-person Hiroshima AI Process Friends Group Meeting, Tokyo, 28 February 2025.
Parliamentary Vice-Minister for Foreign Affairs IKUINA Akiko at the closing ceremony of the 1st in-person Hiroshima AI Process Friends Group Meeting, Tokyo, 28 February 2025. Source: Ministry of Foreign Affairs of Japan.
Closing ceremony of the 1st in-person Hiroshima AI Process Friends Group Meeting, Tokyo, 28 February 2025.
Closing ceremony of the 1st in-person Hiroshima AI Process Friends Group Meeting, Tokyo, 28 February 2025. Source: Ministry of Foreign Affairs of Japan.

MOFA press release, 28 February 2025 · MIC: Results of the Friends Group Meeting

 

OECD HAIP Reporting Framework

Launched on 7 February 2025 as the operational mechanism through which organizations report on their alignment with the HAIP Code of Conduct's eleven actions. The Framework was developed in collaboration with stakeholder communities, incorporating input from government, the private sector, academia, civil society, and research institutions.

v2.0 update (March 2026): The OECD has announced that the HAIP Reporting Framework is being streamlined and improved. A v2.0 pilot is open for new submissions.

Links: transparency.oecd.ai · About the Framework · Submitted reports

CAIDP Statements on HAIP

Statement on AI Policy for Democratic Nations — 2025 G7 Summit Meeting (2 June 2025)

SIGNED BY MERVE HICKOK (CAIDP PRESIDENT) AND MARC ROTENBERG (CAIDP FOUNDER)

CAIDP welcomed HAIP as critical progress while urging the G7 to translate voluntary HAIP principles into enforceable rights-protecting law.

The establishment of the G7 Hiroshima AI Process (HAIP) marked critical progress in shaping international standards. Several AI companies have already submitted HAIP reports in 2025, a meaningful step toward greater public accountability and transparency on internal AI governance. — CAIDP, 2 June 2025
Voluntary commitments will not provide the legal clarity needed to protect human rights as AI-enabled technology becomes ever more entwined with individuals, communities, and nations. We urge G7 leaders to establish enforceable policies to translate HAIP principles into safety and rights-protecting legal frameworks. — CAIDP, 2 June 2025

CAIDP recommendations to the G7 Leaders at Kananaskis:

  1. Ratify the Council of Europe Framework Convention on AI, human rights, democracy, and the rule of law.
  2. Establish clear red lines prohibiting AI technologies that violate human rights, plus a termination obligation for AI systems that evade human control.
  3. Ensure AI standards such as transparency, contestability, and human oversight for all rights-impacting AI systems.
  4. Adopt the recommendations of the G7 Data Protection Authorities.

CAIDP G7 Summit Statement, 2 June 2025 (PDF)

Statement for the G7 Data Protection and Privacy Authorities Roundtable (12 June 2025)

CAIDP SUBMISSION TO THE G7 DPA ROUNDTABLE, HOSTED BY CANADA

CAIDP commended the G7 DPAs for their foundational contribution to HAIP, and connected HAIP transparency to enforceable user rights and to Data Free Flow with Trust (DFFT).

CAIDP commends the G7 DPAs for its foundational work in contributing to the Hiroshima AI Process along with other digital ministers, with its crucial emphasis on meaningful AI transparency to support human and organizational accountability. — CAIDP, 12 June 2025
The G7 Leadership and the Hiroshima AI Process have committed to “the free flow of data while protecting privacy, security, and intellectual property rights” under the umbrella term, Data Free Flow with Trust, with implementation guided by the G7 DPAs and the Institutional Arrangement for Partnership (IAP). — CAIDP, 12 June 2025

CAIDP recommendations to the G7 DPAs:

  1. Establish clear red lines prohibiting AI systems that breach data protection laws or exploit personal data in contravention of human rights.
  2. Endorse a termination obligation for AI systems that evade human control.
  3. Enforce the principles in the G7 DPAs' Statement on AI and Children.
  4. Clarify enforceable standards for transparency, contestability and human oversight for all rights-impacting AI systems.
  5. Clarify standards for Privacy Enhancing Technologies (PETs) within Data Free Flow with Trust (DFFT).

CAIDP G7 DPA Statement, 12 June 2025 (PDF)

 

CAIDP Analysis

HAIP is also covered in CAIDP's AI and Democratic Values Index 2025 and in CAIDP's broader G7 work.

Cross-links: CAIDP AI Index 2025 · CAIDP G7 resource page · CAIDP Council of Europe AI Treaty resource page

Hiroshima AI Process — Documents of Achievement

 

Maintained by the Ministry of Internal Affairs and Communications of Japan.

2023 (Japan)

  • G7 Leaders' Statement on the Hiroshima AI Process — 30 October 2023
  • Hiroshima Process International Guiding Principles for Organizations Developing Advanced AI Systems
  • Hiroshima Process International Code of Conduct for Organizations Developing Advanced AI Systems
  • Hiroshima AI Process G7 Digital & Tech Minister's Statement — 1 December 2023
  • Hiroshima AI Process International Guiding Principles for All AI Actors
  • G7 Leaders' Statement — 6 December 2023

2024 (Italy)

  • G7 Industry, Technology and Digital Ministerial Declaration and Annex 3 “Advancing the Outcomes of the Hiroshima AI Process” — 15 March 2024
  • G7 Joint Statement — 15 October 2024

2025 (Canada)

  • G7 2025 Industry, Digital and Technology Ministerial Declaration — 9 December 2025
  • G7 Industry, Digital and Technology Ministerial Statement on the SME AI Adoption Blueprint Toolkit

Documents of Achievement page (MIC)