TIMEWELL
Solutions
Free ConsultationContact Us
TIMEWELL

Unleashing organizational potential with AI

ISO/IEC 27001 (ISMS) certification mark (SGS / ISMS-AC)

ISO/IEC 27001:2022 certified Certificate No. JP26/00000255 Scope: Planning, development and operation of SaaS products utilizing AI technology

Services

  • ZEROCK
  • TRAFEED
  • TIMEWELL BASE
  • WARP
  • WARP 1Day
  • WARP NEXT Corporate
  • WARP BASIC
  • WARP ENTRE
  • Alumni Salon
  • WARP for Schools
  • Alumni Ventures
  • FDE
  • AI Consulting
  • EMC GLOBAL

Company

  • About Us
  • Team
  • Why TIMEWELL
  • News
  • Contact
  • Free Consultation

Content

  • Insights
  • Knowledge Base
  • Case Studies
  • Whitepapers
  • Events
  • Solutions
  • AI Readiness Check
  • ROI Calculator

Legal

  • Privacy Policy
  • Manual Creator Extension
  • WARP Terms of Service
  • WARP NEXT School Rules
  • Legal Notice
  • Security
  • Anti-Social Policy
  • ZEROCK Terms of Service
  • TIMEWELL BASE Terms of Service

Newsletter

Get the latest AI and DX insights delivered weekly

Your email will only be used for newsletter delivery.

© 2026 株式会社TIMEWELL All rights reserved.

Contact Us
HomeTRAFEED MediaBuilding a Research Security Function: The Owner, the Rules and the Forms, in 90 Days
TRAFEED Media

Building a Research Security Function: The Owner, the Rules and the Forms, in 90 Days

Published2026-07-24Updated2026-11-22Checked against sources2026-11-21Ryuta Hamamoto
research securityresearch integritygovernanceinternal rulesuniversitiesTRAFEED

"We have not done anything yet. Where do we even start?" That is the question I hear most often when research security comes up.

Building a Research Security Function: The Owner, the Rules and the Forms, in 90 Days
Share

(November 22, 2026) Two corrections. We had said measures such as application restrictions apply where declared information turns out to be false; following Section 2-6 of the manual, they apply to intentional false declarations, concealment and similar breaches, in light of their seriousness and consequences. We had also described the three items in Section 3-2 as things the manual wants ready at the point of application; the manual says it is desirable to prepare for putting the arrangements in place, and that the institution may keep those three items in mind in doing so.

(November 22, 2026) We had written that "every statement about training in the manual" is "desirable." We have corrected this: the statements aimed at institutions and researchers are "desirable," while the Cabinet Office's preparation of training materials and delivery of training is "necessary," to be done as appropriate (Part II, Chapter 1, 1-1). We also changed who replied at the 6th expert meeting to "the chair or secretariat," as the minutes summary labels it.

Hello, this is Ryuta Hamamoto from TIMEWELL.

"We have not done anything yet. Where do we even start?"

That is the question I hear most often when research security comes up. Explainers about the regime have multiplied. What nobody has written down is where an institution with nothing in place should put its first step. The procedures manual will tell you that something is "necessary." It will not tell you what to do on Monday morning.

So this piece is about sequence. Who to name as the owner and which department should hold the work, how far to go with internal rules and forms, how to keep the checklist cycle turning, and which parts you do not need to carry in-house. If the regime itself is new to you, start with the complete guide to research integrity and research security and come back here.

Two things get misread more than anything else, so let me put them first. One, the manual does not ask for zero risk. It limits the technologies in scope and then asks for reasonable handling proportionate to the degree of risk1. You do not need to brace yourself for "we cannot apply until the system is perfect." Two, the manual explicitly prohibits discriminatory treatment on grounds of nationality, race, religion, culture or the like1. MEXT takes the same position2. Blanket exclusion of researchers from a particular country or background sits outside what the regime was designed to do. Share those two points with your team at the outset and the later discussions hold their shape much better.

One more thing worth checking early. How far your export control arrangements have already been built is part of the foundation research security sits on. If you want a rough read on where you stand, our free export control self-check is a reasonable place to start.

Two things set the order: the foundation, and the deadline

Why begin with sequence at all? Two reasons.

The first is that the regime itself is built in layers. In its December 2024 policy direction, MEXT wrote that in ensuring research security, "as its foundation, it is first important that researchers themselves and universities and other institutions thoroughly carry out efforts to ensure research integrity, including securing transparency in research activities and autonomous risk management"2. The Cabinet Office manual is constructed the same way, stating that measures to ensure research security are built on top of the thorough implementation of the research integrity work institutions have been doing all along1.

Let me untangle the two terms, because they get used interchangeably and they are not the same thing. Research integrity is the set of understandings and behaviours expected of institutions and researchers to keep research activity sound, fair and transparent, and it applies to all research activity. Research security protects research activity from risks threatening national and economic security, and applies only to the research the state or the institution has judged needs protecting1. Broad foundation, narrow structure on top. Once you see that relationship the order of work decides itself: you cannot bolt a research security procedure onto an institution with no working integrity declaration cycle underneath it. Integrity work is not new either. The starting point is the policy adopted at the Integrated Innovation Strategy Promotion Council on 27 April 2021.

The phrase "Specified Research and Development Program" recurs throughout this article, so here is what it means. Among competitive research funding programmes that presume publication of results, these are the ones the competent ministry has designated, in consultation with the funding agency, as potentially involving technologies on the list of critical technology areas1. Not every competitive funding programme is in scope. Skip that point and your internal discussion swerves straight into "so do we have to do this for every research project?"

Narrowing what you protect is not a Japanese invention. In Europe and North America the shorthand is "small yard and high fence," or "as open as possible, as closed as necessary," and in June 2022 the G7 set out common values and principles on research security and integrity, one of which is "shared responsibility": no organisation can address research security alone, and every member of the research community should recognise its own role and responsibilities. Limiting the scope in order to protect an open research environment is the international common understanding, which is useful language to have when you explain this internally.

The second reason is the deadline. The Cabinet Office set April 2026 as the start of operation, and at the seventh meeting of the expert panel the secretariat stated it would "proceed with concrete measures towards the start of operation in April 2026"3. ATLA's Innovative Science and Technology Initiative for Security became a Specified Research and Development Program from the FY2026 call, with applications open from 13 March 2026 to noon on 20 May. The selection results published on 20 October show 47 projects selected from 560 applications4. JST's Strategic Basic Research Programs apply to projects newly selected in FY20265. For any institution that intends to apply, the day the call opens is the real deadline.

Get the scope right here. In JST's FY2026 application guidelines, what was designated is the five CREST research areas only. PRESTO and ACT-X are outside the scope5. The five cover quantum science and technology, innovative ultra-long-life materials, real-environment intelligent systems, an interdisciplinary systems foundation for human-AI coexistence, and the fusion frontier of light with information, communications, sensing and materials. Institutions in scope are the "universities etc." and "companies etc." named in the commissioned research contract5. Read the programme title alone, conclude that all Strategic Basic Research is in scope, and you have put an unnecessary load on the whole campus. Read it too loosely the other way and you miss an area that does apply.

The risk management the regime asks for turns in four stages: risk identification, risk assessment, risk mitigation measures, and follow-up1. The people in scope are the principal investigator, the representative on the collaborating institution side (Co-PI), and research participants belonging to the lead institution, which includes students1. Building a function means deciding who turns those four stages, in what order, and leaving a trail that can be followed afterwards.

Here is the awkward part. The real risk management work starts once you are shortlisted, but the preparation has to be finished before you apply or it does not land in time. How the requirement came to be embedded as a condition of application is covered in Research Security as a Funding Condition.

The first decision is the owner and the department, not the rules

The first thing to do is not to write internal rules. It is to decide who owns the work.

The manual, in fact, does not specify a department. What it asks you to designate are "persons involved in the operation of the Specified Research and Development Program," which is a set of people, not a box on the org chart. Line up the surrounding documents, though, and an outline appears. The Cabinet Office Q&A on the manual (updated 26 August 2026) asks which unit should serve as the "department for preventing technical information leakage" that carries risk management. Its answer: the units already handling security export control or research integrity would likely be at the centre, and an institution without a suitable unit needs to designate one, including through reorganisation if necessary6. MEXT's notice on the advice desk asks that enquiries "be made centrally through the responsible department within the institution"7. And JST's FY2026 application guidelines say questionnaire responses are submitted "having obtained confirmation from the responsible department at the institutions of the principal investigator and the main co-investigators (including, where a department with responsibility for research security and research integrity has been established, that department)"5.

Nothing there says it is compulsory. Even so, the fact that a funding agency's application guidelines have started constructing sentences on the assumption that such a department exists carries weight. At a minimum, seen from outside, there has to be one place people know to ask.

So should you create one? There is primary data on this. In the supplementary questionnaire to the Cabinet Office's FY2025 follow-up survey, the most common approach was expanding an existing organisation. Among national research and development agencies, 26% created something new, 48% expanded an existing organisation, and 26% set up a committee or task force of departmental heads. Across 32 research institutions the split was 22%, 56% and 22%8.

I think that result is the practical answer. Security export control, conflict of interest management, industry-academia collaboration, HR, research integrity. The information research security needs is already held by those departments. What is missing is not an organisation, it is the lines connecting them. Build from scratch and you spend your first months re-establishing pipes to the departments that hold the data. In the same survey, on whether the institution had "put in place an organisational structure capable of appropriate management," 277 of 344 national, public and private universities had already done so, and adding the 23 planning to within the year takes it to 300, or 87% (the Cabinet Office summary rounds this to "about 90%")8. Starting from nothing now puts you in the minority.

Replace siloed classification work with AI.

METI's FY2024 data shows 52% of foreign exchange law violations stem from classification errors. Download the TRAFEED product catalog covering features and rollout.

Download Product CatalogContact About TRAFEED

For national research and development agencies: start from the add-ons

If you work at a national research and development agency, read this article's premise a little differently. In the Cabinet Office's FY2025 follow-up survey, all 23 national R&D agencies reported implementing all seven items (building understanding among stakeholders, rules on conflicts of interest and commitment, an organisational structure, risk management rules, a mechanism to verify facts objectively, a mechanism to act before risks materialise, and a mechanism to receive updated information). The per-institution table shows all 23 answering "implemented since previous years" on every item89. National R&D agencies are not starting from zero.

Those seven items, however, are research integrity items. The manual adds procedures on top of them for institutions applying to Specified Research and Development Programs. That add-on is where a national R&D agency starts.

Add-on How the manual treats it What the Q&A and other sources clarify What to check at your agency
Arrangements to manage the information needed for risk management Necessary; a structure in which related units collect it together and manage it centrally is desirable1 The units handling security export control or research integrity would likely be at the centre; without one, designate one, including through reorganisation6 Which existing risk management committee or export control unit owns it
Persons involved in operation, and due diligence (13 items) PI, Co-PI and research participants; items 3–8, 10 and 11 cover the past three years1. The Acquisition, Technology & Logistics Agency's FY2026 call (questionnaire, Attachment 5) keeps "the past three years" in the text of section 3(1) and labels the boxes for the same items "FY2023 to present"10 "Three years" means the fiscal year containing the application (or appointment) date plus the two before it — for an appointment on 1 June 2026, FY2024–FY20266. In the Acquisition, Technology & Logistics Agency's FY2026 call (13 March to noon on 20 May 2026), an application made in March gives FY2023–FY2025, which matches the boxes; for one made from April onward, the boxes are one fiscal year wider. Where the boxes state a period, collect for that period10. Administrative staff supporting the application are "involved in operation," but due diligence covers researchers only6 Which items HR records can fill and which need the researcher's own declaration
Self-declaration, attestation and consent When receiving personal information, the institution asks for a consent form on purpose of use and third-party provision, and an attestation that the declaration is current and free of falsehood or omission (3-3, reference)1 The institution checks the attested declaration against public information for inconsistencies. HR data on education and career may be used if the researcher confirms it is current. The lead institution obtains the Co-PI's declaration, consent and attestation (normally through the collaborating institution's administration)6 The route for collecting forms from collaborators when you are the lead institution
Scope of declarations Support other than research funding, patent applications, foreign talent recruitment programmes and more Support other than funding is not limited to money (fees, salaries, scholarships); the QA says it is considered to include offers of posts and provision of goods. The specific scope to declare follows the institution's own rules and judgement. Patents: those filed within three years counting from the filing date, including unpublished and personally held filings (anonymised submission is acceptable). Foreign talent programmes: the manual has in mind the list under the U.S. National Defense Authorization Act6 Items and worked examples in your declaration form
Institution-level checks on collaborators and supporters Desirable. The collaborating institutions covered are those other than Japanese universities, inter-university research institutes, colleges of technology, national R&D agencies and public testing and research institutes (3-3(1)(iii))1 That means all foreign institutions, private companies in Japan or abroad, non-profits and foundations. Kept at "desirable" to limit the burden6 Collaborations with other national R&D agencies or Japanese universities are outside it; set the procedure for companies and foreign institutions
Four items in contracts and agreements Checking the appropriateness of cooperation content, access to research data, handling of IP, and confidentiality is necessary for both lead and collaborating institutions1 — Whether your joint research contract template covers all four
Information on data management (routine) Desirable. The institution collects information on data management plans (DMPs), on how its information systems are managed, and on how its facilities and equipment are managed (3-1(3))1 — Which units hold the information on DMPs, information systems, and facilities and equipment
Data management for the research (at application) Desirable. The lead institution checks the appropriateness of the research's DMP and of the management policies for samples and other material that cannot be managed electronically, for the information systems used (including those the PI manages), and for the facilities and equipment used (buildings, offices, laboratories, experimental apparatus and so on). The same applies to collaborating institutions, reading PI as Co-PI (3-3(1)(ii), (2)(ii))1 — Whether each project has a DMP and all three management policies
When technology has (or may have) leaked — Whether or not the manual applies, report as quickly as possible to those concerned, including the funding agency6 A contact route that works at night and on holidays

Spread across 90 days, the nine add-ons might look like this. The owners and deadlines are our suggestion, not something the manual sets. The evidence column is the record you will use when answering the funding agency's questionnaire and, if something happens, when explaining how your arrangements were set up (manual, Part I, 2-6).

Add-on Owner (example) Deadline (example) Evidence to keep
System for managing information Research promotion office (may double as the export control office) Day 30 The chair's decision designating the department; where the information lives and who can see it
Persons involved in operation, and due diligence Research promotion office + HR Forms by day 60; one dry run by day 90 Designation records; 13-item check records (source, date, checker)
Self-declaration, attestation and consent Research promotion office (legal reviews the forms) Day 60 Form versions and approval dates; collection log
Scope of declarations Research promotion office Day 60 Worked examples in the declaration form (offers of posts and provision of goods, patents, foreign talent programmes)
Institution-level checks on collaborators and supporters Industry-academia collaboration office Day 60 Procedure for partnering with companies and foreign institutions; list-screening records
Four contract and agreement points Industry-academia collaboration office + legal Day 60 Revised contract template; contract review checkbox
Information on data management Research promotion office + information systems + facilities management Day 60 A list of where the information on DMPs, information systems, and facilities and equipment is kept
Data management for the research Research promotion office (together with the PI) One dry run by day 90 Records of checking a project's DMP and its management policies for samples, information systems, and facilities and equipment
When technology has (or may have) leaked Research promotion office + information security Contacts by day 30; tabletop exercise by day 90 Contact flow covering nights and holidays; exercise record

Where to go for advice also differs from universities. The MEXT research security advice desk accepts enquiries from universities, junior colleges, inter-university research institute corporations and colleges of technology; national R&D agencies are not included7. They consult the Cabinet Office's one-stop desk for the manual and the funding agency they are applying to. Section 1-2 of the manual makes it necessary for the government to set up a desk that takes enquiries and requests for advice on the manual in one place, and says it is appropriate for that desk to pass them on to the right body according to their content, for example by referring questions about how a specific call is run to the funding agency running it1. The Cabinet Office says it has set up this desk and lists it on a page alongside its Q&A. It takes enquiries by email and phone, and the email address is given in notices from the ministries concerned. The page does not limit which institutions may use it11. Funding agencies, for their part, must respond appropriately to institutions' questions and requests for advice on how a call is run (Section 2-4(1) of the manual)1. The Q&A likewise answers questions on substituting the checklist, collecting information by electronic form, or using a list designated by a partner country by pointing to the funding agency6.

National R&D agencies have a second face as funding agencies. Of the 21 funding agencies in the same survey, five were national R&D agencies with competitive research funding programmes in FY20258. The Cabinet Office supports funding agencies' arrangements through its FY2026 hands-on support programme; five of the seven organisations approved (AMED, JST, NARO's Bio-oriented Technology Research Advancement Institution, NICT and JAXA) are national R&D agencies12. Part II of the manual also treats funding agencies (Chapter 2) and research institutions (Chapter 3) in separate chapters. My recommendation is not to mix, in one register, the declarations you collect from your researchers to apply and the information you receive about other institutions in order to review. Personal information gathered for different purposes becomes hard to account for once it is mixed.

As an applicant, a national R&D agency counts as a "public research institution" in ATLA's programme, which defines that term as incorporated administrative agencies (including national R&D agencies), special corporations and local incorporated administrative agencies. In FY2026, public research institutions won 10 commissioned and 1 subsidised projects4. JST's guidelines define eligible institutions as "universities etc." or "companies etc." under the commissioned research contract5; check your contract for which applies.

A fillable version of the 90 days for national R&D agencies: the add-on table, the 90-day table, the draft headings for the 13 articles of the rules (with their basis), and the five-step emergency contact flow, laid out in Excel with owner, deadline and status columns. → 90-day research security plan for national R&D agencies (Excel, in Japanese)

For the article-by-article mapping between the export control rules and the research security rules, and the steps up to the president's decision, see One Rulebook for Export Control and Research Security at Japan's National R&D Agencies.

Ninety days. What to put where

From here I will lay this out in 90-day blocks. To be clear, the 90-day frame is not from any government document. It is a sequence I put together from the conversations I have been asked into, based on what the manual treats as necessary and on working backwards from the call schedules. Month one, decide the container. Month two, produce the paper. Month three, run it once. It is the four stages of risk management mapped onto your own calendar.

The first 30 days: measure where you are, decide the container

Three things in the first month.

Measure where you are. The yardstick is Annex 2 of the manual, the "Research security checklist." It comes in three versions: for the lead institution, for collaborating institutions and for researchers. Each asks whether you can carry out what the manual requires when applying to a Specified Research and Development Program, answered "can / cannot" and the like1. Section 2-5(2) of the manual says it is desirable for applicants and their institutions to self-assess with this checklist, and the QA adds that doing so routinely, not only when an application is in view, is also desirable6.

The lead-institution version asks about the 13-item due diligence, checking the four contract and agreement points, submitting information on your own and collaborating institutions (such as financial standing and whether they appear on lists; Japanese universities, inter-university research institute corporations, colleges of technology, national research and development agencies and public research institutes are excluded, item 1(3) of Annex 2), notifying Co-PIs' institutions, consent and attestation forms, risk mitigation measures, and follow-up after the research starts (adding participants, reporting errors in declarations)1. Every item you mark "cannot" becomes work for day 31 onwards. Whether you may partly substitute public information or your own researcher directory, or collect information through an online form, is something the QA says to settle "while consulting the funding agency in charge of the programme, after checking the call for proposals," so ask the agency you are applying to6.

The seven questions in the Cabinet Office follow-up survey (organisational structure, mechanisms for objectively verifying facts, mechanisms for receiving updates to information, and so on) still work as a supplementary yardstick for the research integrity foundation8. National research and development agencies are the exception: in the FY2025 survey all 23 reported implementing all seven items8. Scored this way, they will find no gaps, so they should use Annex 2 and the table of add-ons in the section on national R&D agencies above.

Fill in the Annex 2 self-check directly: an Excel workbook that lists the questions of the three Annex 2 checklists word for word and moves every row marked "partly" or "cannot" onto the 90-day plan sheet. It also has a due diligence log for the 13 items and a draft-answer sheet for the questionnaire that works out the three fiscal years from the application date. → Research security self-check and due diligence kit (Excel, in Japanese)

Decide the container. Take stock of existing departments and list who holds what information. Then designate the responsible department and name both an executive-level owner and a working-level owner. What matters is fixing the names, and it is fine if both roles are held alongside other duties. Start drafting rules while the answer to "who owns this?" is still "general affairs, probably" and you will stall halfway through. I have watched it happen.

For a national research and development agency, check the existing internal control framework before building anything new. The Act on General Rules for Incorporated Administrative Agencies requires the statement of operational procedures to cover the development of systems to ensure that officers' duties are performed in compliance with laws and regulations and other systems to ensure that operations are proper (Article 28(2))13. Deciding where the research security owner and responsible department sit within that framework (a risk management committee or a compliance department, for example) saves you from adding another committee. Confirm the reporting line to the competent ministry at the same time; it makes the funding agencies' questionnaires easier to answer. Settle how the rules are adopted, too. The head of the agency (the president) represents it and presides over its business (Act on General Rules, Art. 19(1)). Changing the statement of operational procedures needs the competent minister's approval (Art. 28(1)), so keeping only the outline of the system there and putting the designation of officers and the forms in internal rules adopted by the president's decision keeps the process light (check each agency's rule-making procedures for which rules go through which process)13.

Write down your external route. MEXT's research security advice desk was established by notice 7-Monkasho-Ka No. 54, dated 10 April 2025, issued jointly by the Director-General of the Science, Technology and Innovation Policy Bureau and the Director-General of the Higher Education Bureau. It is open to universities, junior colleges, inter-university research institute corporations and colleges of technology. The notice sets conditions: enquiries should be routed centrally through the responsible department, and should come after the institution has done some internal analysis and found itself unsure how to proceed7. Without an internal escalation route, in other words, you cannot use the desk at all. Responses to information security incidents are outside its scope.

Days 31 to 60: put the rules and forms on paper

Month two is documents. What you need to produce is more clear-cut than people expect.

Start with the research integrity checklist. The Cabinet Office and MEXT publish templates in two versions, one for researchers and one for institutions. The institutional template asks whether you have an advice desk; whether education and training opportunities exist; whether you have a mechanism to identify and address information where the institution has risk concerns; whether there is a procedure for the institution to verify and decide when signing memoranda and similar documents; whether you assess risk on the counterparty organisation and its members against the End User List and comparable lists in other countries; whether you have a mechanism to receive reports of subsidies, remuneration or goods provided by foreign institutions; and whether you can see the content and purpose of long or frequent overseas travel14. Write your internal rules so that you can answer yes to those questions and the design work goes much faster.

The manual makes it "necessary" for institutions to collect the information researchers declare on the basis of the checklist, and "necessary" to review the checklist in light of how it has been operating1. Producing it once and filing it away is not what was contemplated. Write the annual declaration cycle, and how you receive updates in between, into the rules themselves.

If you are drafting research security rules from scratch, matching the article headings to the manual makes it easier to reconcile them with the funding agency's questionnaire. The following is our suggestion; adding a chapter to existing research integrity or export control rules works just as well.

Article Heading (suggestion) Where it sits in the manual1
1 Purpose Part I, 1-1
2 Definitions (Specified Research and Development Program, persons involved in operation, PI, Co-PI, research participants, due diligence, lists) Reference (1), definitions
3 Responsible officer and department Part II, 3-1(2)
4 Declaration and central management of research activity information Part II, 3-1(1)(2)
5 Designation of persons involved in operation and their confidentiality Part II, 3-2 (1)(2)
6 Self-declaration, consent and attestation Part II, 3-3 (reference) on personal information
7 Due diligence and risk assessment Part II, 3-3(1)(2)
8 Review of joint research contracts and agreements (four points) Part II, 3-3(1)(v), (2)(iii)
9 Risk mitigation measures Part II, 3-4
10 Follow-up after the research starts (adding participants, reporting errors) Part II, 3-5
11 Emergency contact and response Part II, 3-2 (3); Part I, 2-6
12 Training Part II, 3-1(4)
13 Record retention and review of the rules and checklist Checklist review: Part II, 3-1(4). Record retention is our proposal (the manual sets no retention period)

Next, the three forms: the self-declaration form, the consent form, and the written declaration. The consent form covers two things. That the researcher declares personal data to the institution for the purpose of risk management and that the institution uses it only for that purpose, and that the institution may provide it to third parties for that purpose1. Third parties here means the funding agency and relevant government bodies, plus the lead institution where the declaration is made to a collaborating institution. The written declaration carries one sentence: that the personal data declared is, to the signatory's knowledge, current as at the time of declaration and contains no false content and no omissions1. When designing the forms, tabulate who you collect what from, split by PI, Co-PI and research participant. Students count as research participants, so work out in advance who in each lab is in scope.

Then add the contract review angle. For joint research contracts and other agreements the manual makes it "necessary" to verify the appropriateness of four things: the content of the cooperation, access to research data, the handling of inventions, patents and other intellectual property, and the terms of confidentiality1. Due diligence on the collaborating institution itself only reaches "desirable," but those four contract points are in the necessary column. In most institutions this means adding review criteria to an existing contract flow. High priority, low cost.

Put the security export control (FEFTA) checks into the same contract review and you avoid doing the work twice. The four research security points are required by the manual and the export control checks by FEFTA, so the legal bases differ, but the documents and the people overlap. The pairing below is our own.

Research security (the manual's four points) Export control checks for the same contract
Content of the cooperation Whether the technology or goods provided fall under the list controls in the Foreign Exchange Order or Export Trade Control Order (classification), and whether the public-domain or basic-science exceptions apply
Access to research data End-user and end-use checks and screening against the End User List; whether participants are non-residents or fall under the specified categories
Handling of intellectual property Plans to publish results (relevant to the public-domain exception)
Terms of confidentiality Limits on subcontracting and disclosure to third parties, no use outside the purpose, return of data and samples at the end

The ten pre-contract checks and sample clauses are in Joint and commissioned research with overseas institutions: ten checks before signing.

Last, compress your emergency escalation and response flow onto a single page. The manual says that, at the point of application, it is desirable for the institution to prepare for putting its risk management arrangements in place, and that in doing so it may keep three things in mind: designating the persons involved in running the programme, making sure those persons observe their confidentiality obligations, and arrangements for contact and response in an emergency such as leakage of critical technology (3-2)1. That part is written as "desirable" and "may." Elsewhere in the same manual, though, verifying whether those arrangements were appropriately in place and functioning effectively, once critical technology has leaked, is "necessary"1. The distinction came up at the expert panel, and the secretariat explained that the intention is not to require completed arrangements at the point of application3. Turn that around and the real message appears: after something happens, the state of your arrangements will be examined. One page of escalation flow up front strikes me as the rational move.

Five steps are usually enough (our suggestion):

  1. Whoever notices (researcher or administrator) contacts the research security department; at night and on holidays, via the information security on-call team (a CSIRT or similar)
  2. The department reports to the executive-level owner at once and decides how to stop the damage spreading (suspending access, isolating devices and so on)
  3. The owner reports to the funding agency as soon as a possible leak is recognised. The QA says reporting as quickly as possible to those concerned, including the funding agency, is necessary whether or not the manual applies6
  4. If the Foreign Exchange and Foreign Trade Act may have been breached, the export control officer decides on reporting to the Minister of Economy, Trade and Industry (Ministerial Ordinance on Compliance Standards for Exporters, Article 1, item 2 (nu))
  5. Record what happened and what was decided, so that you can later verify whether the arrangements worked (manual, Part I, 2-6)

The manual is equally direct on where responsibility sits. Intentional false declarations, concealment and similar breaches of the manual can, in light of how serious they are and the gravity of what they led to, be treated as improper receipt under the "Guidelines on the Appropriate Execution of Competitive Research Funding," and can trigger measures such as restrictions on future applications1. (Section 2-6 of the manual puts this in the future tense: it says such measures "could" be taken after the Guidelines are amended. Calls for designated programmes, such as JST's, already state that application restrictions may be imposed.) On the other side, it states that where technology has nonetheless leaked despite sufficient measures having been taken, the research institution and the researcher do not bear responsibility1. Put those two sentences side by side and the purpose of the whole exercise becomes visible. It is not to reduce incidents to zero. It is to be able to explain, afterwards, that reasonable measures were taken. Measures with no record behind them are not material for that explanation, which is exactly where forms and approval trails designed to leave a history earn their keep.

Days 61 to 90: run it once

Month three is about putting what you built into motion. The institutions I worry about most are the ones that assemble the paperwork and stop there, so do not skip this.

Run one dry run of due diligence. Due diligence is the process of verifying the appropriateness of institutions and researchers participating in the research, and the version conducted using public information alone is called open source due diligence, or OSDD1. The manual's 13 verification items are these.

# Item to verify Past three years
1 Academic background
2 Research and employment history
3 Research funding received Yes
4 Support other than research funding received Yes
5 First, corresponding and co-authorship on published papers Yes
6 Patent filing status Yes
7 Participation in foreign talent recruitment programmes Yes
8 Disciplinary history under the guidelines Yes
9 Whether the person appears on a list
10 Whether the person belongs to a listed institution Yes
11 Whether the person has relationships with researchers at listed institutions Yes
12 Status as a "non-resident" or under a "specific category" in export control
13 Anything else the funding agency deems necessary

Items 3 through 8, plus 10 and 11, cover the past three years including the fiscal year in which you apply1. Not every item runs three years back, so keep that straight when you design the form.

The scope sounds heavy, but the manual does not raise the bar on where information has to come from. Due diligence may be carried out using "self-declared information, open source information and other information that each lead institution can ordinarily obtain"1. The tools it gives as examples are all public: academic paper databases, portals such as Google Scholar, researcher databases such as e-Rad and researchmap, IP databases such as J-PlatPat, METI's End User List, and the United States consolidated screening list1. Pick one researcher and fill in all 13 items to the end within that range. You will find out exactly where you get stuck. How to gather the material item by item is covered in Due Diligence in Practice and How to Answer the Questionnaire.

Before you go live, share one thing internally. Appearing on any country's list is a regulatory designation. It is not a judgment about the organisation listed, or about the researchers who work there. The same goes for "specific categories" in security export control. The framework has applied since May 2022, and METI explains that "specific categories are simply a categorisation of cases requiring individual verification during screening, and falling under a specific category does not mean the person is regarded as posing a security concern" (p. 3 of that document; we checked the original on November 21, 2026)15. Falling under one is not a reason to exclude anyone. You verify, and if a procedure is needed you follow it. Treating people differently on grounds of nationality or origin is, as noted at the top, prohibited by the manual itself1. Distribute a checklist without explaining any of this and you generate a lot of avoidable tension on the ground.

Alongside that, try answering the questionnaire you would submit to the funding agency, as if today were the day. Some items will be unanswerable. That set is your remaining task list.

Running one round of training is worth doing too. Both statements about training aimed at institutions and researchers sit in the "desirable" column: institutions are encouraged to recommend training to researchers intending to apply, and researchers are encouraged to take it proactively. The Cabinet Office's own work, preparing training materials and running training, is something it "needs to" do as appropriate1. At the 6th meeting of the expert panel (23 October 2025), a written comment from an absent member, introduced by the secretariat, argued for upgrading this to "necessary"; the chair or secretariat (the summary does not say which) replied that "desirable" would stay because the programmes do not yet exist16.

Which means the off-the-shelf material available today is thin. The manual does position the production of training material and the holding of briefing sessions as work for the government to do1, but what can actually be inspected right now is roughly the research integrity near-miss casebook produced under a FY2024 MEXT commissioned project, plus the checklist templates mentioned earlier. MEXT's policy direction sets out three lines of work, supporting forums for inter-university collaboration, establishing an advice desk connected to relevant government bodies, and developing training material and programmes2, with pilot activity from FY2025. My view is that running one session yourself with the casebook and the checklists beats waiting for the material to be finished.

Note too that the scope reaches research participants including students. What that means for how labs take people on is set out in Taking On Students and RAs Under Research Security.

Work out what you do not need to carry in-house

Run the 90 days and you will always be left with a piece that people cannot do by hand. Let me close on that.

External resources first. From 19 February to 24 March 2025, the Cabinet Office ran an open call for its programme supporting risk management arrangements for research security and integrity, and selected eight institutions, whose names were not published17. The expert panel minutes note that under the FY2024 supplementary budget, funding went to 23 universities and research institutions3. For FY2026 a hands-on support programme aimed at funding agencies is running, with seven organisations approved: AMED, JST, JSPS, NARO (Bio-oriented Technology Research Advancement Institution), ERCA, NICT and JAXA12. Support menus for research institutions change shape year to year, so check the Cabinet Office and MEXT pages periodically. For universities, MEXT ran an open call for its FY2026 "University Advisory Hub Programme for Research Integrity and Research Security" (16 July to 20 August 2026, three proposals) and selected two: the University of Electro-Communications (with the University of Tsukuba as joint proposer) and Osaka University18. It is a commissioned project to build a platform where universities share their practices and know-how.

That leaves the largest remaining load, the due diligence work itself, and item 11 in particular: whether the person has relationships with researchers at listed institutions. The manual defines the relationship as a record of joint or commissioned research, co-authored papers, or named co-presentations at conferences and similar events1. Over three years, you compare the affiliations of the person's direct partners (joint research partners, co-authors and co-presenters) with the listed institutions. Name matching alone does not finish the job, because you need each partner's affiliation at the time. The manual acknowledges this, noting that where adequate due diligence is difficult using the listed public tools alone, using commercial information analysis tools or commissioning investigations from firms may also be considered1.

Even when you use outside tools or commission work, the final determination belongs to the research institution. What you can hand off is assembling the material; keeping its reasoning in a form you can explain afterwards stays with the institution. The centralised management the manual calls desirable, meanwhile, does not require merging departments; aligning where information lives and how it is reached gets you there.

Also readJapan's national R&D agencies: connecting information security to research security under note 33How do you share data management work with the information security team at a national R&D agency?

What to take away

  • The regime sets the order. A working research integrity declaration cycle is the foundation, and the research security procedure sits on top of it
  • The manual has been in operation since April 2026. The next marker is the opening day of the calls you intend to apply to. Preparation has to be finished before the call opens
  • National research and development agencies have all implemented the seven research integrity items. Their starting point is the add-ons in the manual (due diligence, consent and attestation forms, the four contract items, data management, reporting a leak), and they consult the Cabinet Office's one-stop desk for the manual and the funding agency
  • Scope is limited to Specified Research and Development Programs. At JST that covers the five CREST research areas only; PRESTO and ACT-X are outside it. Check before designing anything campus-wide
  • The manual does not ask for zero risk. Where technology leaks despite sufficient measures, the institution and the researcher do not bear responsibility. The goal is being able to explain what you did
  • The people in scope are the PI, Co-PIs and research participants including students. Appearing on a list, or falling under a specific category, is a regulatory designation and not a judgment about the organisation or the person
  • The first move is not internal rules. It is naming the owner and the responsible department, and the data shows expanding an existing organisation beats creating a new one
  • Measure where you are with Annex 2 of the manual, the research security checklist (lead institution, collaborating institution and researcher versions). Whatever you mark "cannot" is the work for month two onwards
  • Month two is paperwork: the checklist revision, the self-declaration form, the consent form, the written declaration, the four contract review points, and the emergency escalation flow
  • Month three is the dry run. Push one case all the way through and your remaining gaps take concrete shape
  • Centralised management only reaches "desirable," but it is worth designing early

"Building a governance function" sounds like a large undertaking. Lay out what the 90 days actually contain and it comes down to fixing some names, producing four kinds of document, and running the thing once. The hard part is not the content. It is the first step of deciding whose job this is. The expert panel discussed the acute shortage of specialist staff and the need to train them3, and building from a position of not having the people is the same condition every institution faces.

So start with one name. It can be someone who already has a full role. Write it down this week, and the rest of the 90 days has something to attach itself to.

If you want due diligence and centralised information management to run as a routine process with sources and records kept, you can contact us about TRAFEED.

References and primary sources

Footnotes

  1. Expert Panel on Ensuring Research Security and Research Integrity, "Procedures Manual for Ensuring Research Security," December 2025 (Cabinet Office) https://www8.cao.go.jp/cstp/kokusaiteki/integrity/yushikisha/guidelines_v1.pdf ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12 ↩13 ↩14 ↩15 ↩16 ↩17 ↩18 ↩19 ↩20 ↩21 ↩22 ↩23 ↩24 ↩25 ↩26 ↩27 ↩28 ↩29 ↩30 ↩31 ↩32 ↩33 ↩34 ↩35 ↩36

  2. MEXT, Science, Technology and Innovation Policy Bureau, "Direction of Concrete Measures on Research Security in MEXT-Related Programmes for Universities and Other Institutions," 18 December 2024 https://www.mext.go.jp/content/20241218-mxt_kagkoku-000039402_1-1rrr.pdf ↩ ↩2 ↩3

  3. Cabinet Office, "Summary Minutes of the Expert Panel on Ensuring Research Security and Research Integrity (7th Meeting)," 1 December 2025 https://www8.cao.go.jp/cstp/kokusaiteki/integrity/yushikisha/7kai/gijirokugaiyo.pdf ↩ ↩2 ↩3 ↩4

  4. Ministry of Defense, "FY2026 selection results of the Innovative Science and Technology Initiative for Security" (in Japanese), 20 October 2026. https://www.mod.go.jp/atla/pinup/pinup_r081020.pdf . Checked 20 November 2026. ↩ ↩2

  5. Japan Science and Technology Agency (JST), "FY2026 Strategic Basic Research Programs (CREST, PRESTO, ACT-X) Application Guidelines" https://www.jst.go.jp/kisoken/boshuu/teian/koubo/2026youkou.pdf ↩ ↩2 ↩3 ↩4 ↩5

  6. Cabinet Office, "Q&A on the Procedures Manual for Ensuring Research Security" (in Japanese), last updated 26 August 2026. https://www8.cao.go.jp/cstp/kokusaiteki/integrity/security_qa_v1.pdf . Checked 20 November 2026. ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12

  7. MEXT, "Establishment of the MEXT Research Security Advice Desk (Notice)," 7-Monkasho-Ka No. 54, 10 April 2025 https://www.mext.go.jp/content/20250424-mxt_kagkoku-000039402_1.pdf ↩ ↩2 ↩3

  8. Cabinet Office, "Summary of FY2025 Research Integrity Follow-up Survey Results," December 2025 https://www8.cao.go.jp/cstp/kokusaiteki/integrity/ri_follow-up_fy2025/ri_fu_fy2025_sum.pdf ↩ ↩2 ↩3 ↩4 ↩5 ↩6

  9. Cabinet Office, "FY2025 Research Integrity Follow-up Survey Results (Research Institutions, by Institution)" (in Japanese). https://www8.cao.go.jp/cstp/kokusaiteki/integrity/ri_follow-up_fy2025/ri_fu_fy2025_ri_re.pdf . We counted the answers of all 23 national R&D agencies. Checked 20 November 2026. ↩

  10. Acquisition, Technology & Logistics Agency, "FY2026 Innovative Science and Technology Initiative for Security (commissioned projects) call for proposals" (in Japanese), March 2026, Attachment 5 "Questionnaire on research security," the text of 3(1) and the boxes for items 3–8, 10 and 11; call period in 4.3.2 (13 March 2026 to noon on 20 May 2026). https://www.mod.go.jp/atla/funding/koubo/r08/r08koubo_full_itaku.pdf . Checked against a copy obtained on 20 November 2026. ↩ ↩2

  11. Cabinet Office, "Q&A on the manual and the one-stop desk for the manual" (in Japanese) https://www8.cao.go.jp/cstp/kokusaiteki/effort/contact.html . The page says the Cabinet Office has set up an enquiry desk for the Procedures Manual for Ensuring Research Security (known as the one-stop desk for the manual), and that the desk's email address is given in notices from the ministries concerned. Checked on 22 November 2026. ↩

  12. Cabinet Office, "Hands-on support programme for funding agencies' research security risk management arrangements: list of approved organisations" (in Japanese). https://www8.cao.go.jp/cstp/kokusaiteki/integrity/r8_supportdecision.pdf . Checked 20 November 2026. ↩ ↩2

  13. e-Gov Laws, "Act on General Rules for Incorporated Administrative Agencies" (Act No. 103 of 1999), Article 19(1) and Article 28 (statement of operational procedures), paragraphs 1 and 2 (in Japanese). https://laws.e-gov.go.jp/law/411AC0000000103 . Checked 20 November 2026. ↩ ↩2

  14. MEXT, "Checklist Template for Universities and Research Institutions," version of 29 June 2023 https://www.mext.go.jp/content/20230704-mxt_kagkoku-000019002_1.pdf ↩

  15. METI, Trade Control Department, Security Export Control Policy Division, "On the Clarification of Deemed Export Management" https://www.meti.go.jp/policy/anpo/law_document/minashi/meikakukanitsuite2.pdf (7 pages). Page 3 states that the specific categories are "simply a typological grouping of cases requiring individual verification during screening," and that falling under one "does not mean the person is regarded as posing a security concern." Original checked (November 21, 2026). The same caveat also appears in a reference sheet on the specific-category approach posted by Tokyo Metropolitan University (May 2022): https://bunnkei-kanribu.jim.tmu.ac.jp/site/wp-content/uploads/2023/08/%E3%80%90%E5%8F%82%E8%80%83%E8%B3%87%E6%96%991%E3%80%91%E2%91%A4%E7%89%B9%E5%AE%9A%E9%A1%9E%E5%9E%8B%E3%82%A2%E3%83%97%E3%83%AD%E3%83%BC%E3%83%81%E3%81%A8%E3%81%AF-2022.5%EF%BC%88%E6%97%A5%E3%83%BB%E8%8B%B1%EF%BC%89.pdf ↩

  16. Cabinet Office, "Summary Minutes of the Expert Panel on Ensuring Research Security and Research Integrity (6th Meeting)," 23 October 2025, "Questions from absent members (introduced by the secretariat)" item 5, https://www8.cao.go.jp/cstp/kokusaiteki/integrity/yushikisha/6kai/gijirokugaiyo.pdf (in Japanese; checked 21 November 2026) ↩

  17. Cabinet Office, "Open call for the programme supporting risk management arrangements for research security and integrity" (in Japanese). https://www8.cao.go.jp/cstp/kokusaiteki/integrity/kobo_r7.html (call period 19 February to 24 March 2025; selection result: eight institutions selected). Checked 21 November 2026. ↩

  18. MEXT, "Selection results for the FY2026 University Advisory Hub Programme for Research Integrity and Research Security" (in Japanese) https://www.mext.go.jp/content/20261007-mxt_kagkoku-000019002_1.pdf . Original checked 21 November 2026. The programme's purpose is taken from MEXT's "Research Integrity and Research Security" page (https://www.mext.go.jp/a_menu/kagaku/integrity/index.html , checked the same day). ↩

This article was produced with the help of AI. A human verified the primary sources and edited the text before publication.

Read next

Articles that answer the question you are likely to have next.

  1. 1Implementing Risk Mitigation Measures: How Far to Go with the Seven Examples in Research SecurityOnce the structure is in place, how far should mitigation measures go?
  2. 2Japan's national R&D agencies: connecting information security to research security under note 33How do you share data management work with the information security team at a national R&D agency?
  3. 3Research Security Training in 30 Minutes: Japan's Manual, a 10-Question Quiz and RecordsHow a Japanese university or research institution can build a 30-minute research security training session. Where training sits in the Cabinet Office's Procedures Manual (Part II 3-1(4), 3-4 and 4-3: reviewing the checklist is "necessary", encouraging training is "desirable"), who to train and when, a 30-minute outline, a 10-question quiz with sources, and how to connect attendance records to item 4(1) of ATLA's questionnaire (Annex 5) and to post-start checks. Checked against the original documents.

This article is part of TRAFEED Media, our practical media on export controls and economic security

Back to TRAFEED MediaThis week's regulatory updates

52% of FY2024 export-control violations stem from classification errors. Is your team covered?

METI FY2024 data shows over half of violations stem from classification. Start with a free 5-question light check (~2 min, no email), then continue to the full 10-question report.

Contact About TRAFEED
Contact About TRAFEEDContact form (about 3 min)TRAFEED product briefFeatures & rollout in PDF

Share this article if you found it useful

Share

Newsletter

Get the latest AI and DX insights delivered weekly

Your email will only be used for newsletter delivery.

Free download

Recommended materials

Research Integrity & Security Self-Check Sheet (for universities and research institutions, 2026)

A fill-in self-check covering disclosure, the 13 due-diligence items, and the 4-step risk management, based on the Cabinet Office "Procedures for Ensuring Research Security" (December 2025) and the Integrated Innovation Strategy Promotion Council policy.

China MOFCOM Lists (Control & Watch): 80-Entity Reference Sheet (through the 29 Jun 2026 wave)

All 80 entities on China MOFCOM's Control List and Watch List, in official / Chinese / English names. Being listed is a regulatory designation, not a judgment of the company — useful both for counterparty screening and for listed companies to understand where they stand. Based on primary sources (Announcements No. 11/12/27/28); as of 20 November 2026 there had been no additions or removals (checked through No. 56).

The BIS Affiliates Rule (50 Percent Rule) Readiness Checklist (2026)

A fill-in working checklist for trade compliance teams preparing for the BIS Affiliates Rule. Covers what the rule actually does (aggregation across owners and across lists, cascade/carry-forward, the rule of most restrictiveness, and why exactly 50% is in while below 50% is not automatically out), what is stayed and what is in effect now, where ownership data typically breaks, and a six-phase preparation list from scoping through governance. The stay was due to end on 9 November 2026; on that day BIS published a rule extending it to 10 January 2027 (91 FR 70789). The sheet is built to work whether the rule then returns, is stayed again, or is modified (as of 21 November 2026). Based on primary sources (90 FR 47201, 90 FR 50857, 91 FR 70789, 15 CFR, and China's Ministry of Commerce). Designation is a regulatory classification, not a judgement about a company; final determinations rest with your export-control officer and the authorities' current guidance.

See all materials

Related Knowledge Base

Export Control Guide

Solutions

Strengthen Export ComplianceStreamline compliance with complex export regulations

Talk with us about export-control operations

Share your screening, classification, or compliance workflow. We will map where TRAFEED can help—via our contact form (no cold booking).

Contact UsDownload Catalog

Related Articles

What Is Research Security? Japan's Procedures Manual Went Live in FY2026, and the 13 Due Diligence Checks It Requires

What Is Research Security? Japan's Procedures Manual Went Live in FY2026, and the 13 Due Diligence Checks It Requires

When I talk with university faculty and research administration staff, the questions have changed noticeably over the past six months.

2026-07-24
Research Integrity and Economic Security at Universities: A Practical Guide to Deemed Exports and Research Security (2026 Edition)

Research Integrity and Economic Security at Universities: A Practical Guide to Deemed Exports and Research Security (2026 Edition)

A practitioner's overview of how research integrity and research security have been built into Japan's economic security policy.

2026-07-19
Research Integrity and Research Security: A Complete Guide to the Regime That Went Live in FY2026

Research Integrity and Research Security: A Complete Guide to the Regime That Went Live in FY2026

A single map of research integrity and research security, built from the regime that actually started operating in FY2026.

2026-07-24
Foreign Funding Disclosure in Japan: 10 Questions on Outside Appointments and Cross-Appointments

Foreign Funding Disclosure in Japan: 10 Questions on Outside Appointments and Cross-Appointments

Ten questions on disclosing foreign funding, outside appointments, cross-appointments and emeritus titles in Japanese grant applications, answered from the original rules, with sample entries.

2026-11-20
Undisclosed Funding and Dual Affiliation: Reading the Research Security Verification Items Through the Regime That Produced Them

Undisclosed Funding and Dual Affiliation: Reading the Research Security Verification Items Through the Regime That Produced Them

Undisclosed foreign funding, dual affiliation, participation in foreign talent recruitment programmes, unintended technology leakage, insider risk.

2026-07-24
One Rulebook for Export Control and Research Security at Japan's National R&D Agencies

One Rulebook for Export Control and Research Security at Japan's National R&D Agencies

How a Japanese national research and development agency can bring its security export control rules and its research security rules into one framework. I compare the strength of the three sources (FEFTA and the compliance standards ordinance, METI's fifth-edition guidance, and the Cabinet Office procedures manual), show what belongs in the statement of operational procedures and what belongs in internal rules (Article 28 of the Act on General Rules), map the articles side by side, set out the path to the president's approval, and explain why records held as a funding agency should stay separate.

2026-11-22