Now
1. Brazil Proclaims a Result and Leaves Its Orders Unreferended 34,393votes cast after 5:00 pm under Brazil’s extension, 0.027% of turnout and far below the margin 2. The First Committee Clears All Four Military-AI Texts £490mwhat Table 6.1 of Britain’s Budget scores for AI over six years, against "over £700 million" in the text 3. Egypt's Grace Period Has Expired. The Regulator Says It Starts Now 13 Novembernew close of India’s 10,000-GPU tender, moved by an unannounced corrigendum 4. The FCC Has Not Ruled on AI Robocalls, and Says It Has No Plans To 77.9%turnout in Israel’s final count, published 30 October 5. Mexico's Supreme Court Parks 138 Biometric-CURP Appeals 163–5First Committee vote to "advance to negotiations" on autonomous weapons 6. Canada and the EU Write AI Conformity Into a Summit Text $25,000New York’s per-violation penalty for unsafe AI companion features, from 1 January 2027 7. South Australia's Royal Commission Opens With Six Themes and a Carve-Out 243→16South Korean local governments folding into sixteen regional platforms by end-2029 8. Beijing Certifies 21 More Personal-Information Auditors 0.478F1 of a clinical de-identifier on real nursing notes, against 0.994 on its synthetic benchmark 9. Islamabad Chooses Age Assurance Over a Ban 180,000Japanese officials being moved onto the Digital Agency’s agent tooling 10. New York's Courts Tell Lawyers How to Check What AI Wrote $0.47what a US federal agency pays for Gemini for Government per year, until 15 November Washington Writes ‘Super Intelligence’ Into the Rulebook $3,000Pennsylvania’s maximum civil fine under Act 99, from 21 December Brazil’s Platform Orders Outlive the Vote 6 hourshow often Meta and X must report reach to Brazil’s electoral court Israel’s Count Is Done; the Integrity Argument Is Not £456mUK AI security, procurement and capability commitments, to be confirmed at the Budget India’s 10,000-GPU Tender Quietly Slips to 14 November 2 Decemberwhen the EU nudification ban and the marking duty become enforceable Three Legislatures, One Chatbot Problem 14MWHumain’s Nvidia Blackwell Phase 1, from Q4 2026 Chile Keeps Boric’s AI Bill Alive While Kast Reserves a Rewrite 700disinformation complaints logged by South Africa’s Electoral Commission The Baselines Are the News: What Britain’s Red Book Scores for AI 27 OctIsrael votes under its first deepfake-disclosure law A Covenant and a Queue Jump: How the UK Will Pick Its Data Centres 54risks listed in China’s AI Safety Governance Framework 3.0, up from 30 €30bn+EU call for up to seven AI gigafactories 554deepfakes logged in Brazil’s first round; only 371 labelled 30+AI firms sent the first AI Act information requests 164AI-made ads in the US midterms, by one count 42signatories to Canada’s voluntary data-centre principles 174MWCape Town data centre now under appeal 83 dayswait reported under Medicare’s AI pilot, against a 72-hour standard 20home-grown foundation models backed by the IndiaAI Mission 2%of Turkish public investment budgets earmarked for AI
AI&GPP
Sign in Subscribe

Tokyo and Berlin Link Their AI Safety Institutes

Japan and Germany have agreed to link their AI safety institutes, open a high-level AI dialogue and coordinate on a Japan-hosted AI Summit. The pledges sit inside an economic-security declaration aimed at critical-mineral coercion; none of the nine memoranda on Tokyo's list covers AI.

Security & Geopolitics desk
A low, glass-fronted modern government building with a flat overhanging roof, seen across a lawn with stone paving, with office towers behind it and a Japanese flag on a pole to the left.
The east entrance of the Prime Minister’s Office in Tokyo, where Takaichi Sanae and Friedrich Merz met on 28 October 2026. Photo: Cabinet Public Affairs Office, Cabinet Secretariat of Japan, CC BY 4.0, via Wikimedia Commons.

Prime Minister Takaichi Sanae and Chancellor Friedrich Merz met for about 30 minutes in Tokyo on 28 October 2026, according to the Ministry of Foreign Affairs’ record, and issued three texts: a Japan-Germany Joint Leaders’ Statement, a Joint Leaders’ Declaration on Economic Security Cooperation and a Joint Leaders’ Declaration on Advanced Technology Cooperation. For AI policy the operative line is paragraph 18 of the statement: the leaders “took note of recent AI related incidents and concurred on the collaboration between Japan AI Safety Institute and German AI Safety and Security Institute”. Germany’s institute is nine weeks old.

Where the AI commitments are, and what form they take

The substance is in the economic-security declaration, not the technology one. Its “Artificial Intelligence” section welcomes “the establishment of a high-level dialogue on AI” with a first meeting “at the earliest opportunity”, and undertakes to explore cooperation on “highly capable AI models and light AI models” through “our AISIs, namely the Japan AI Safety Institute (J-AISI) and the AI Safety and Security Institute Deutschland (AISI Deutschland), with particular attention to the misuse of AI against critical infrastructure”. It affirms “access to and collaboration on frontier models among trusted partners” and welcomes Japan’s announcement that it will host an AI Summit.

Document (28 October 2026)What it says on AIForm
Joint Leaders’ Statement, paras 16 and 18AI named among digital-connectivity priorities; AISI collaboration agreed; “recent AI related incidents” noted but not identifiedLeaders’ statement
Declaration on Economic Security Cooperation, section 5High-level dialogue; J-AISI and AISI Deutschland cooperation on highly capable and light models and critical infrastructure; frontier-model access among trusted partners; AI supply chains; Global South; Japan-hosted AI SummitLeaders’ declaration
Declaration on Advanced Technology CooperationAI listed by cross-reference only; no AI measuresLeaders’ declaration
Nine memoranda on the Ministry of Foreign Affairs’ list (seven between the governments, two with companies)None on AISigned instruments

None of this is a treaty. The Ministry of Foreign Affairs’ list of memoranda signed or exchanged around the visit runs to nine items, from a semiconductor partnership to the protection of nationals abroad, two of them between a ministry or agency and a company; none mentions AI. The institute link rests on two leaders’ texts and on what the institutes do next. According to the German federal government’s transcript of the joint press appearance, Merz spoke of more than 20 agreements between companies and Takaichi of around 30 at state and private level.

Two institutes at very different stages

J-AISI dates from February 2024. Its July 2026 overview places the secretariat inside the Innovation Platform Agency (IPA), with 12 ministries and agencies and five related organisations around it, and lists guides on red-teaming, evaluation and incident response. AISI Deutschland was founded on 31 August 2026 by the Federal Ministry for Digital and State Modernisation (BMDS) and the Federal Ministry of the Interior, implementing a June 2026 National Security Council decision, the BMDS announced. It begins as a “virtual” nucleus: the Federal Office for Information Security (BSI) takes security, the Federal Network Agency (BNetzA) takes safety, the first job is assessing leading models’ cyber capabilities and risks, and a second phase has no date. Berlin counts France, the United Kingdom and the EU AI Office among its existing partners and says it will join the international network of safety institutes; J-AISI’s overview of that network, now the International Network for Advanced AI Measurement, Evaluation and Science, lists ten members, and Germany is not among them. The Tokyo pledge links an established evaluator to one still being built.

Why safety sits inside an economic-security text

The surrounding paragraphs explain the placement. Paragraph 19 of the statement records “grave concern over the increasing use of economic coercion such as, arbitrary export restrictions and retaliatory measures on critical minerals”; the declaration lists “industrial AI” and robotics among the supply chains on which the two governments will explore cooperation. Neither text names China as the source of that coercion. Mainichi reported that Tokyo had Beijing’s tightening export controls in mind, that the dialogue is aimed at joint AI development and that Merz travelled with about 30 executives. Read beside Beijing’s decision to treat model weights as a controlled export, the clause on frontier-model access “among trusted partners” is as much a supply-security promise as a safety one, in keeping with the two-bloc pattern in AI governance.

Where it fits Tokyo’s dialogue-building

Germany follows France, which agreed a leaders’-level AI dialogue with Japan in April 2026; its first session met in Paris on 19 June 2026, with an intergovernmental segment on national AI capability, the Global South and dual-use deployment and a public-private segment with ten companies. Both dialogues feed Japan’s plan to host an AI Summit. At home, the same government has just published a government-wide operating standard for AI agents and has asked for a far larger AI budget for the next fiscal year; the German link gives that programme another G7 evaluator to compare notes with.

What to watch

Whether the paragraph becomes practice will show on three dates: the first AI dialogue, due “at the earliest opportunity”; the next economic-security consultations, due before the end of 2026; and the second Japan-Germany Intergovernmental Consultations, brought forward to Berlin in 2027. Until the institutes publish a joint evaluation or a shared method, procurement and security teams should read the link as convergence in testing, not mutual recognition of results.

Sources

  1. Japan-Germany Joint Leaders’ Statement — Ministry of Foreign Affairs of Japan, 28 October 2026 (PDF)
  2. Japan-Germany Joint Leaders’ Declaration on Economic Security Cooperation — Ministry of Foreign Affairs of Japan, 28 October 2026 (PDF)
  3. Japan-Germany Joint Leaders’ Declaration on Advanced Technology Cooperation — Ministry of Foreign Affairs of Japan, 28 October 2026 (PDF)
  4. 日独首脳会談 — 外務省 (Ministry of Foreign Affairs of Japan), 28 October 2026 (in Japanese; translation by AI & GPP)
  5. Pressestatement des Bundeskanzlers und Japans Ministerpräsidentin Takaichi — Bundesregierung, 28 October 2026 (in German; translation by AI & GPP)
  6. AISI Deutschland gegründet — Bundesministerium für Digitales und Staatsmodernisierung, 31 August 2026 (in German; translation by AI & GPP)
  7. Japan AI Safety Institute (J-AISI): Overview — Japan AI Safety Institute, 1 July 2026 (PDF)
  8. 1st Japan-France High-Level Dialogue on AI — Ministry of Foreign Affairs of Japan, 20 June 2026
  9. 日独首脳、「AI対話」創設で合意　対中念頭に重要鉱物でも協力 — 毎日新聞 (Mainichi Shimbun), 28 October 2026 (in Japanese; translation by AI & GPP)
  10. 2026年10月のメルツ独首相訪日を契機に日独政府間及び政府と民間セクター間で署名・交換された覚書等一覧 — 外務省 (Ministry of Foreign Affairs of Japan), October 2026 (PDF; in Japanese with English document titles; translation by AI & GPP)

AI & GPP reports on how artificial intelligence and automation are changing the way governments decide, regulate and campaign. Corrections and tips: contact the editors.

Get the stories that matter to decision-makers, weekly.

Discussion

No comments yet. Start the conversation.

Discussion is open to members.

Join free Sign in

Read next

The Baselines Are the News: What Britain’s Red Book Scores for AI

HM Treasury's Budget text says 'over £700 million' for AI. Its own scorecard books £490m of new borrowing-funded spend over six years and £160m of Sovereign AI financial transactions, back-loaded to 2028-29, with the savings it is meant to unlock asserted rather than itemised.

163 to 5: New York Backs Talks on Autonomous Weapons

The UN First Committee has cleared all four military-AI texts: 163 to 5 for 'advance to negotiations' on autonomous weapons, 171 to 4 for two years of Geneva dialogues, a consensus CCW omnibus. What the counts settle before the Review Conference, and what they cannot.

Gulf AI: One Institution, Three Memoranda, One Number

In the week to 25 October 2026 the Gulf's AI news arrived as paperwork: a ratified Qatari decision creating a national AI centre, three memoranda from Oman's state visit to Britain, and one Saudi figure — 14MW — that can be checked on a date.