Hello, this is Ryuta Hamamoto from TIMEWELL.
"Next year one of our faculty will apply to a Specified Research and Development Program. We heard we should run research security training. What should it cover, who should take it, and how long should it be?" Suppose that is the question a research administration or research integrity office at a Japanese university gets from a faculty member. My short answer: Japan's Procedures Manual treats training as "desirable", but one 30-minute session before you collect self-declarations cuts down on missed items and follow-up questions, and it gives you a record if you later choose training as a mitigation measure. That is the order I recommend.
The short version
- For institutions and researchers, training is "desirable" in the manual (Part II 3-1(4) and 4-3). The same paragraph 3-1(4) makes reviewing the research integrity checklist "necessary". Producing training materials and running training is "necessary" for the Cabinet Office, which is to do so as appropriate (Part II 1-1).
- Build the 30 minutes in this order: purpose and the ban on discrimination, the four steps, the 13 verification items, missed items and errors, one case and where to get advice, and a quiz. Train the PI, the Co-PI and research participants, students included, and extend it to the administrative staff who support the program.
- Training is one of the example mitigation measures (Part II 3-4). On ATLA's questionnaire (Annex 5), it is an option in 4(1), which applies when you answer "yes" to 3(2). Record the date, the versions of the materials and the manual, the quiz result, and the due diligence record it links to.
This article does not assess any country, institution or person. Being on Japan's End User List or the U.S. Consolidated Screening List is a regulatory designation, not a judgment on the listed institution or on the researchers who work there. The manual itself says that, in judging whether a partner can be trusted, "naturally, there must be no discriminatory treatment on the grounds of nationality, race, religion, culture or the like" (Part I 1-1(2); our translation)1. I treat that sentence as part of the training content.
I checked the facts against copies retrieved on 22 November 2026 of the Cabinet Office's "Procedures Manual for Ensuring Research Security" (December 2025), its Q&A (last updated 26 August 2026), the one-stop desk page, the explanatory slides, and the expert panel's meeting summaries123456. I read ATLA's FY2026 call for proposals (commissioned research) in a copy retrieved on 20 November 20267. All of these are in Japanese; quotations are our translations. The manual reuses section numbers across chapters (there is a "2-2" in Part I and another in Part II), so I always give the part and chapter, as in "Part II chapter 3, 3-1(4)". Governance as a whole is covered in setting up research security governance, and the pre-application check in the Annex 2 self-check. Here I stick to building the training and keeping the record.
Training is "desirable": how strong the manual's words are, and what the panel said
The manual defines its own words. Part I 1-2(2) says that where the text says "necessary", it describes the minimum measures to be implemented, and where it says "desirable", it describes measures that are "desirable to implement, but whose implementation is not required at present"1. The Cabinet Office desk page confirms it: asked whether everything not marked "desirable" is mandatory, it answers, "That is correct"3. That is the yardstick for reading what the manual says about training.
The body of the manual mentions training in four places.
| Where | Whose measure | What the text says | Strength |
|---|---|---|---|
| Part II 1-1 (i) | Government (Cabinet Office) | Producing research security training materials, holding briefings and running training | Necessary (to be done as appropriate) |
| Part II 3-1(4) | Research institution | When its researchers apply, or plan to apply, to a Specified Research and Development Program, encourage them and others to take research security training | Desirable |
| Part II 3-4 | Research institution | "Improving research security literacy through training" is one of the example risk mitigation measures | Mitigation itself is necessary; training is one example |
| Part II 4-3 | Researcher | Actively take training to improve research security literacy | Desirable |
Watch the two sentences of 3-1(4). The first says it is "necessary" for the institution to review its research integrity checklist in light of how it is working. What gets reviewed is the research integrity checklist, not the training. Only the second sentence, on encouraging training, is "desirable"1. The Cabinet Office's explanatory slides mark the first "necessary" and the second "desirable" in the 3-1 table (slide 26)4. If an internal memo sums it up as "3-1(4) is necessary", readers will think training is mandatory too, so keep the two apart.
3-4 reads a little differently. After reviewing and assessing risk, an institution must ("necessary") take risk mitigation measures. The measures are then described as "examples such as the following", and training is one of the seven examples. The text adds that "reasonable measures proportionate to the level of risk are sufficient"1. So training is not itself an obligation. It is one tool you can pick once you decide mitigation is needed.
Why it stayed "desirable" is on the record. In the summary of the sixth expert panel meeting (23 October 2025), item 5 of the "questions from absent members (introduced by the secretariat)" proposes that researcher training should read "researchers expected to work on a Specified Research and Development Program ... need to take" it. The same item says it needs to be examined who would supply the training (or a model for it). The response: "Since the programs are not yet in place, we will keep the expression 'desirable'"5. In the summary's legend, the response mark stands for "the chair or the secretariat". Earlier, at the fifth meeting (18 July 2025), a member said it was unrealistic to collect information on every staff member at all times, and the response was that the plan was a two-stage approach that would "first aim to raise literacy"8.
At the seventh meeting (1 December 2025), the government side said briefings would be held once the manual was finalized. Because foreign researchers in Japan who do not understand Japanese are also covered, it said the government would consider producing an English translation on its own responsibility. It also said support would continue through "producing training materials and running training", as the manual states6. MEXT's policy direction of 18 December 2024 likewise says it will "develop and roll out training materials and training programs" on research integrity and research security9.
Research ethics education makes a useful contrast. JST's FY2026 call makes completing a research ethics education program an application requirement for the applicant (section 6.2; for CREST, co-investigators do not have to have completed it at the time of application)10. ATLA's FY2026 call requires every researcher on a commissioned project to take compliance education and research ethics education during contracting and to understand the content, and it asks institutions to submit completion reports (2.6.3(2))7. In the manual and the two calls I read, research security training has no such application requirement or completion report.
So where are the government's materials? When I checked the Cabinet Office's "Measures for ensuring research security" page on 22 November 2026, it listed the manual, the list of critical technology areas, the explanatory slides, the Q&A and the desk, the expert panel, and the support programs. I did not see anything published as training materials11. I could not find an English version of the manual on the Cabinet Office's English page either. Q&A II-1 says, about foreign-language versions of the manual, "we plan to produce an English version first"212. Rather than wait, I think it is faster to build one 30-minute session yourself from the manual and the explanatory slides. When the government's materials come out, swap them in.
Who to train, and when
Start from the people the manual puts within risk management. Part I 2-4 names three groups for the lead institution: the PI, the Co-PI, and research participants belonging to the lead institution. It defines research participants as "researchers who plan to take part in the Specified Research and Development Program (including students, and excluding the PI and Co-PI)". A collaborating institution must also cover its Co-PI and its own research participants1. Students came up at the sixth panel meeting, where the record says students taking part in the research are within due diligence, and that this is "understood as necessary"5.
| Who | Where in the manual | What to stress | Suggested timing (our proposal) |
|---|---|---|---|
| PI | Part I 2-4 (i). In the researcher checklist (Annex 2), the PI prompts participants to submit information and confirms it has all been submitted | The full 30 minutes, plus prompting participants and checking submissions | When you decide to apply, before collecting self-declarations |
| Co-PI | Part I 2-4 (ii). The Co-PI's own collaborating institution also runs due diligence | The full 30 minutes, plus sending documents to both their own institution and the lead institution | Same. The collaborating institution runs it; share the materials |
| Research participants (including students) | Part I 2-4 (iii). If not decided at application, add them as soon as they are decided (note 9) | The full 30 minutes, in English as well | When participation is decided |
| Participants added after the research starts | Part II 3-5. Due diligence is required when adding participants | The full 30 minutes, or a recording and the quiz | Before they join |
| Administrative staff supporting the application and program | Part II 3-2 (i)(ii) (naming those involved in running the program and enforcing confidentiality; points the manual says "could be considered"), Q&A II-11 | How the 13 items are handled, personal information, confidentiality | When preparations for the application begin |
The staff row needs a note. Q&A II-11 says that people involved in running a program include administrative staff who support the application and operations, but that due diligence by the institution covers "only the 'researchers' taking part in the program"2. So staff training is not about being screened. It is training for the people who handle the personal information and the research content.
As a rule, train before you collect self-declarations. Part II 3-3 makes it "necessary" for the lead institution to have the 13 verification items self-declared and then checked, and to carry out due diligence, by the deadline the funding agency sets in its call1. Deadlines differ by funder. ATLA's FY2026 commissioned research call ran from 13 March to noon on 20 May 2026, and lead institutions of projects selected for interview had to submit the questionnaire by around mid-July (5.13.3)7. For JST's FY2026 Strategic Basic Research Programs, the CREST call ran from 7 April to noon on 2 June, and the questionnaire was to be sent separately to the principal investigators of projects subject to risk management, with a deadline JST would set (6.5)10. As of 22 November 2026 I have not confirmed when the FY2027 calls will open.
I recommend four points in time. First, once in the fiscal year before a call, for labs that plan to apply; this is the "plan to apply" case in 3-1(4). Second, when a lab decides to apply, before you hand out the declaration forms. This one matters most. Third, at contracting after selection; where a funder already requires research ethics education, as ATLA does, fold it into the same notice. Fourth, whenever a participant joins after the research starts.
If foreign researchers or international students are among the participants, run the same 30 minutes in English. The seventh panel meeting also confirmed that foreign researchers in Japan who do not understand Japanese are covered6. The point of an English session is not to treat foreign nationals differently. It is so everyone covered understands the same content in the same way. The Cabinet Office's English page carries provisional translations of the research integrity checklist templates (the researcher version of 17 December 2021 and the university and research institute version of 29 June 2023)12. When you quote the manual in English, say it is your institution's provisional translation until an official English version appears.
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.
A 30-minute outline: four steps, 13 items, no discrimination, and missed items
Thirty minutes is our own guide from working through the process, not an official standard. I suggest it because it fits one slot of a lab meeting and can be repeated for every application. The order follows where researchers get stuck when filling in their own declarations, not the order of the manual's chapters.
| Time | Topic | What to say | Source |
|---|---|---|---|
| 0:00 to 0:04 | Purpose and no discrimination | This is a check that lets international research continue; zero risk is not required; no judgments by nationality, race, religion, culture or the like | Part I 1-1(1)(2) |
| 0:04 to 0:09 | Four steps and roles | Risk identification, risk assessment, risk mitigation, follow-up. Researchers declare, the institution checks, and the funding agency checks too | Part I 2-3 and 2-4, Part II 4-1 |
| 0:09 to 0:16 | The 13 items and the three fiscal years | The list and which period applies; items 4, 6, 7 and 11 cause the most confusion | Part II 3-3(1)(i), Q&A II-4, II-12, II-14, II-15, II-16 |
| 0:16 to 0:20 | Missed items and errors | Consent form and declaration, who to tell when you find an error, how breaches of the manual are handled | Part II 3-3 (reference box) and 3-5, Part I 2-6 |
| 0:20 to 0:25 | One case and where to get advice | Read one case; confirm the campus contact and the reporting route | MEXT casebook, Q&A I-7, Part II 1-2 |
| 0:25 to 0:30 | Quiz | The 10 questions in the next section | This article |
Say the purpose and the ban on discrimination in the first four minutes
Open with why the process exists. Part I 1-1(1) says its purpose is to let institutions and researchers confirm that partners in international research can be trusted, so they can do their research with peace of mind. Part I 1-1(2) says the manual does not ask for zero risk and asks for reasonable responses proportionate to the risk1. From a researcher's point of view, this is not a process to stop research with other countries. It is a check that lets it continue. Whether you say that up front changes how the next 26 minutes land.
Sample script: "This process is a check that lets us keep doing joint research with partners abroad. The manual does not ask for zero risk. It also says plainly that, in judging whether a partner can be trusted, there must be no discriminatory treatment on the grounds of nationality, race, religion, culture or the like. The 13 items we cover today look at background, funding, affiliations, contracts and track records of collaboration. They are not a way of judging people by nationality."
One addition. ATLA's questionnaire (Annex 5) uses item 13 for "nationality"7. In the manual, item 13 is "other matters the funding agency deems necessary for due diligence", a field each funder defines1. In the session, add one sentence: writing a nationality in a field is different from deciding things based on nationality. MEXT's policy direction says "there must be no discrimination by race, nationality or the like" and that care is needed so Japan's measures are not seen internationally as discriminatory9. On item 12's "specific categories" under Japan's deemed export rules, METI's deemed export Q&A, question 59, says that for license applications "we do not envisage uniformly refusing a license solely because a person falls under a specific category"13.
For the four steps and 13 items, go deep only where people get stuck
In the next five minutes, show the whole flow on one slide. Part I 2-3 names the four steps an institution takes when applying to a Specified Research and Development Program (risk identification, risk assessment, risk mitigation and follow-up) and makes the series "necessary". The funding agency must take the same four steps1. The researcher's role is in Part II 4-1. Researchers who plan to take part must ("necessary"), at the institution's request, declare information on their research activities in addition to the usual research integrity declarations. Self-checking with the research security checklist is "desirable"1.
The 13 verification items are (1) to (13) in Part II 3-3(1)(i). Reading them all out would blow the seven minutes, so put the list on a slide and talk only about the hard parts. Start with the period. Items 3 to 8, 10 and 11 cover "the past three years including the fiscal year in which the application date falls"; the body of the manual puts no three-year limit on items 1, 2, 9, 12 and 131. Q&A II-4 explains three years as the fiscal year of the application date (or appointment date) plus the two fiscal years before it, "three years in total"2. For an application in May 2026, that is FY2026, FY2025 and FY2024 (our own working; Japan's fiscal year runs April to March). If the funder's form states a period, have people follow the form. In ATLA's Annex 5, the fields for items 3 to 8, 10 and 11 say "FY2023 (Reiwa 5) to the present"7.
Four items cause most of the confusion. Item 4, support other than research funding, means remuneration and salary, scholarships, donations, honorary titles and similar, and concurrent positions (the definition in Part II 2-1(2)(iv), repeated in Annexes 2 and 3). Q&A II-12 says it is thought to include not only money but also offers of positions and provision of goods, and that the exact scope should follow the institution's own rules and judgment12. Decide your institution's scope before the session and tell people on the day. For item 6, patents, Q&A II-14 says every application filed within the period must be self-declared, published or not, including patents in the researcher's own name that the institution does not know about2. For item 7, foreign talent recruitment programs, note 22 of the manual gives as an example the programs on the list compiled under section 1286 of the U.S. National Defense Authorization Act for FY2019 (the "1286 list"), and ATLA's Annex 5 asks applicants to list participation in programs on that list "without omission"17. The "relationships" in item 11 are joint or commissioned research, co-authored papers, and joint oral presentations at conferences1. Where a paper has hundreds of co-authors, the desk page says self-declared co-authors are within due diligence, the manual does not ask for zero risk, and if a co-author of concern turns up, the institution should assess the level of risk and, where needed, discuss mitigation with the funding agency3.
Treat missed items as a matter of fixing, not blame
This is the part researchers brace for, so I suggest covering how to fix things before the consequences. First, the paperwork. The reference box in Part II 3-3 says that when an institution receives personal information from researchers, it "shall" ask for a consent form and a declaration. The declaration states that, to the signer's knowledge, the information is current as of the declaration and contains no false statements or omissions1. Q&A II-17 says the institution receives the self-declaration with the declaration attached, then runs due diligence with open sources and checks for discrepancies2.
Next, what happens when someone finds an error. Part II 3-5, on measures after the research starts, says that if an error is found in information a researcher declared, the institution must ("necessary") promptly report it to the funding agency and redo due diligence based on the corrected information1. What you need from researchers is one thing: tell the office as soon as you notice.
Last, breaches. Part I 2-6 says that if the required measures were properly taken and technology still leaked, institutions and researchers are not held responsible. If there is deliberate false declaration or concealment, on the other hand, the government must respond, and it says it is conceivable to revise the funding guidelines so that, depending on how malicious the act was and how serious the result, application restrictions can be imposed as improper receipt of funds1. ATLA's call also says application restrictions may be imposed for acts that breach the manual (5.13.6)7. At the seventh panel meeting, a member pointed out that a third party might leak technology while claiming there was "no intent". The response explained that 2-6 was written to reduce the burden on researchers, and said the existence of such actors, and the need to act at once if you notice such conduct, would be conveyed "through literacy education"6.
Sample script: "If you notice something you left out, contact [the responsible office] that same day. That applies after the research has started, too. The institution reports it to the funding agency and checks again with the corrected information. The manual says that if you took the required steps properly, you are not held responsible even if a leak could not be prevented. What causes problems is deliberate false declaration or concealment."
For consent forms and declarations, you can hand out our research security self-declaration, consent and declaration forms (Word, in Japanese) as they are. They set out the 13 verification items in 3-3 and the required content of the consent form and declaration in the manual's own wording, with a quick table of the three fiscal years and a collection sheet for Co-PI documents.
One case only, paired with where to get advice
Use the remaining five minutes to read one case. For researchers, I find case 5, "Unintended co-authorship", in the researcher section of MEXT's research integrity near-miss casebook (March 2025, produced under an FY2024 MEXT contract) works well14. A partner abroad, as corresponding author, put a paper together. The professor checked only their own part and left the other formalities to the partner, and later found that the co-authors included a researcher from another institution whom the professor had never met or worked with (the casebook replaces university and country names with letters). Among the responses expected of researchers, the casebook lists checking everyone involved in a paper while it is written and published, and talking to the responsible office if there is a concern. It ties straight into items 5 and 11. The casebook predates the manual and is a research integrity resource, so translate its terms into the manual's terms when you use it.
List the campus contact first. Then add the Cabinet Office's one-stop desk for the manual and MEXT's research security consultation desk (set up by a notice dated 10 April 2025)315. The MEXT desk serves universities, junior colleges, inter-university research institute corporations and colleges of technology, and asks that consultations go through one responsible office on campus15. Tell researchers to go to the campus office first. Part II 1-2(2) notes that the Cabinet Office desk advises the person consulting and does not make decisions on their behalf1. This is also where you explain reporting a possible leak. Q&A I-7 says that, whether or not the manual applies, a report to the people concerned, including the funding agency, must be made as quickly as possible, and asks that you report promptly through the contact arrangements you have set up as soon as you recognize that a leak may have occurred2. In the session, say it plainly: contact the campus office when you notice, not after you have confirmed.
A 10-question quiz, with answers and sources
The quiz is ten true-or-false questions. At 30 seconds each, it takes five minutes. Rather than scoring people, look at which questions most people got wrong and use that to fix the notes on the declaration form and your explanation. Every answer rests on the original text of the manual, the Q&A or the desk page.
| No. | Statement | Answer | Source |
|---|---|---|---|
| 1 | The manual requires every researcher who applies to take research security training | False | Part II 3-1(4) and 4-3 say "desirable". Part I 1-2(2) defines "desirable" as not required at present1 |
| 2 | Graduate and undergraduate students who plan to take part in the research are within risk management as research participants | True | Part I 2-4 (iii) defines research participants as "including students"1 |
| 3 | For an application in May 2026, items 3 to 8, 10 and 11 cover FY2026, FY2025 and FY2024 | True | Part II 3-3(1)(i) and the "three years in total" explanation in Q&A II-412. If the funder's form has a period field, follow that too |
| 4 | It is fine to apply extra checks or mitigation to a researcher based only on their nationality | False | Part I 1-1(2): there must be no discriminatory treatment on the grounds of nationality, race, religion, culture or the like1 |
| 5 | A university abroad gave me an honorary title but no money, so I do not need to declare it under item 4 | False | Item 4 includes honorary titles and concurrent positions (Part II 2-1(2)(iv), Annexes 2 and 3). Q&A II-12 adds that offers of positions and provision of goods are thought to be included12 |
| 6 | A patent application filed less than 18 months ago and not yet published does not need to go under item 6 | False | Q&A II-14: applications in the period must be self-declared whether published or not, including ones in your own name that the institution does not know about2 |
| 7 | If you co-authored a paper with a researcher at a listed institution, you cannot apply | False | I found nothing in the manual saying a relationship bars an application. Item 11 records whether a relationship exists, and mitigation is considered based on the due diligence result (Part II 3-3(1)(i) and 3-4). The desk page says to assess the level of risk and, where needed, consult the funding agency13 |
| 8 | The declaration states that, to the signer's knowledge, the information is current at the time and contains no false statements or omissions | True | Reference box on handling personal information in Part II 3-31 |
| 9 | After the research started, I realized I had not declared a concurrent position. I can fix it in next year's declaration | False | Part II 3-5: when an error is found, the institution must promptly report to the funding agency and redo due diligence with the corrected information1 |
| 10 | I think research data may have got out. I will wait until the facts are confirmed before telling the campus office | False | Q&A I-7: report promptly through your contact arrangements as soon as you recognize a leak may have occurred2 |
When you mark the answers, look at where the wrong answers cluster. If many people miss questions 5 and 6, the notes on items 4 and 6 of your declaration form probably need work. If question 4 trips people up, spend longer on the opening four minutes. If anyone marks question 7 as true, that lab may believe "if anything comes up in screening, it's over", which pushes people toward not applying at all. Link each frequently missed question to the document you need to fix, in one line, and fix the form before the next application. I suggest using the quiz less as the end of the training and more as a checklist for the declaration form.
Some labs will find a web form easier than paper for collecting answers. If you go that way, make sure each person's answers can be matched to the same row in the training record described next. For more on question 2, see taking on students and research assistants; on question 5, the Q&A on declaring foreign funding, outside work and cross appointments; and on checking co-authors for question 7, research security due diligence.
Connecting training records to the "training" mitigation option and Annex 5, item 4(1)
To be clear up front: neither the manual nor ATLA's Annex 5 has a provision requiring research security training records to be kept or submitted. That is based on my searches of the full manual and of Annex 5 for the Japanese words for "record", "retain", "attend" and "training"17. I still recommend keeping records, because if you choose training as a mitigation measure, you will later be asked how it was carried out.
In ATLA's Annex 5, training appears in item 4(1). Item 4(1) says that if you answered "yes" to 3(2) under "3. Risk identification and risk assessment", namely "As a result of due diligence, is there anyone for whom risk mitigation measures should be taken?", mitigation measures proportionate to the risk become necessary, and it asks you to choose the measures, more than one if needed. The fifth option is "improving research security literacy through training". Item 4(2) then says "you need to confirm that the answer above is feasible" and has you tick "I have confirmed it is feasible"7. Because 4(1) rests on a "yes" to 3(2), training given when 3(2) is "no" is best treated not as mitigation but as the encouraged training in 3-1(4) (our own reading).
Once a project is selected and the research starts, the record earns its keep. Part II 3-5 makes it "necessary" for the institution to check, as appropriate after the research starts, how mitigation measures are being carried out, and to act on the results. The funding agency must also check implementation after the start (Part II 2-3)1. Whether the mitigation is sufficient is judged by the responsible ministry together with the funding agency, which ask for additional measures if it is not (Part II 1-4 and 2-2(2))1. If you chose training, you cannot answer a check on implementation unless you can show when, who, with which materials, and how understanding was confirmed.
| Field | Example (fictional University Alpha) | Why keep it |
|---|---|---|
| Project and funder | Research on X materials; ATLA's Innovative Science and Technology Initiative for Security (write the fiscal year) | To identify which application the training served |
| Attendee and role | Research participant (doctoral student), name | To match against the groups in Part I 2-4 |
| Date and format | In person at a lab meeting, 30 minutes | To answer post-start checks on implementation (Part II 3-5) |
| Version of the materials | Internal research security training, version 2 | To show later what was taught |
| Versions of source documents | Manual (December 2025), Q&A (last updated 26 August 2026), the funder's call | To find who needs an update after a revision |
| Quiz result | 9 correct, missed question 6 | To feed into fixing the notes on the declaration form |
| Trainer and reviewer | Research administration officer; reviewed by the person in charge of research security | So no single person's judgment closes it |
| Status as a mitigation measure | Chosen under Annex 5, item 4(1)? (yes or no); number of the linked due diligence record | To answer 4(1), 4(2) and post-start checks |
| Non-attendance | Date of the make-up session | So nobody is missed |
If you train a specific person as a mitigation measure, be careful how you record it. Keep the due diligence result and the reasons for the risk assessment in the due diligence record, with access restricted, and put only the linked record number in the training record. If a training roster can be read as a list of who was assessed, it creates needless tension in the lab. Rather than building a special session for particular people, I think it is more measured to take the standard 30 minutes and add an explanation of how the project's data and facilities are managed. Data management is listed in Part II 4-2 as "desirable" for researchers1.
In most situations, I see training alone as a weak mitigation measure. The examples in 3-4 also include managing access to facilities and equipment and managing access to research data, and training often only works in combination with them. Note 31 of the manual says that when mitigation measures are taken, a comprehensive organizational and information management framework needs to be built1. How far to go with each measure is covered in implementing risk mitigation measures.
I found no retention period for these records in the manual. Keep them at least until the research period ends, aligned with how long you keep due diligence records (our proposal). Where a funder already requires completion reports for research ethics education, as ATLA does in 2.6.3(2), adding a research security column to that ledger means one notice and one record7.
To link due diligence records and training records row by row for each person, you can use the "DD record ledger" sheet in our research security self-check and due diligence record kit (Excel, in Japanese). It records one item per row (person by verification item), with the source, date checked, checker, result, risk assessment and whether mitigation is needed, and includes sheets for the Annex 2 self-check and a draft questionnaire answer. For the layout of a training record form, our corporate internal export control training kit (Word, in Japanese) also has one. Its content is export control training, not research security, but its record form, which keeps the versions of the materials and the regulations, can be reused as is.
Summary, and what I have not confirmed
On paper, research security training is "desirable". Even so, one 30-minute session before you collect self-declarations means fewer gaps in the forms and fewer questions later. Cover the purpose and the ban on discrimination, the four steps, the 13 items and the three fiscal years, how to fix missed items and errors, one case and where to get advice, and finish with the ten-question quiz. If you choose training as a mitigation measure, keep the record in a form that answers Annex 5, items 4(1) and 4(2), and the checks after the research starts. If you want one thing to do this week, pick a lab likely to apply next year and try the 30 minutes with the PI and the research participants. The questions most people get wrong are the places to fix on your declaration form.
Here is what I had not confirmed as of 22 November 2026. I do not know when the Cabinet Office will publish training materials. I have not confirmed whether an English version of the manual has been published (Q&A II-1 says an English version "is planned first"). I have not confirmed when ATLA's and JST's FY2027 calls will open or which programs will be designated. The 30-minute outline, the quiz, the record fields and the suggested timing are our own proposals, not official standards. If a funding agency specifies training content or records, follow its instructions. Decisions on risk assessment and mitigation belong to the research institution.
What to read next
- The Annex 2 self-check: before training, check whether your institution can run the procedures.
- The research security questionnaire guide: how to gather the 13 items and answer the questionnaire.
- Implementing risk mitigation measures: how far to go with access controls and the other measures you pair with training.
- Setting up research security governance: owners, rules and forms in 90 days.
- Taking on students and research assistants: contracts and declarations when students join as participants.
- Designing internal export control training: role-based export control training, written for companies.
- The complete guide to research integrity and research security: the overall picture.
You can use TRAFEED for screening against the lists behind items 9 to 11 and for keeping a record of what you checked. To talk it through with us, get in touch here.
Footnotes
-
Expert Panel on Ensuring Research Security and Research Integrity, "Procedures Manual for Ensuring Research Security" (our translation of the Japanese title), December 2025 (Cabinet Office), in Japanese. https://www8.cao.go.jp/cstp/kokusaiteki/integrity/yushikisha/guidelines_v1.pdf . Part I chapter 1, 1-1(1)(2) and 1-2(2); Part I chapter 2, 2-3, 2-4 (note 9) and 2-6; Part II chapter 1, 1-1, 1-2(2) and 1-4; Part II chapter 2, 2-1(2)(iv) (definition of support other than research funding), 2-2(2) and 2-3; Part II chapter 3, 3-1(4), 3-2, 3-3(1)(i) and its reference box on personal information, 3-4 (note 31) and 3-5; Part II chapter 4, 4-1, 4-2 and 4-3; note 22; Annexes 2 and 3. Checked against a copy retrieved on 22 November 2026 (identical to the file retrieved earlier that month). ↩ ↩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
-
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 . I-7, II-1, II-4, II-11, II-12, II-14, II-15, II-16 and II-17. Checked against a copy retrieved on 22 November 2026. ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12
-
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 FAQ entries on measures not marked "desirable" and on co-authors in large collaborations. Checked against a copy retrieved on 22 November 2026. ↩ ↩2 ↩3 ↩4 ↩5
-
Cabinet Office, Science, Technology and Innovation Secretariat, explanatory slides on the Procedures Manual for Ensuring Research Security (in Japanese). https://www8.cao.go.jp/cstp/kokusaiteki/integrity/explanatory_guidelines.pdf . Slide 1 (operation from April 2026) and slide 26 (the 3-1 measures marked "necessary" and "desirable"). Checked against a copy retrieved on 22 November 2026. ↩ ↩2
-
Cabinet Office, summary of the sixth meeting of the Expert Panel on Ensuring Research Security and Research Integrity (held 23 October 2025), in Japanese. https://www8.cao.go.jp/cstp/kokusaiteki/integrity/yushikisha/6kai/gijirokugaiyo.pdf . The legend for the discussion (● expert member, ⇒ chair or secretariat), the response on students and due diligence, and item 5 of the questions from absent members introduced by the secretariat. Checked against a copy retrieved on 22 November 2026. ↩ ↩2 ↩3
-
Cabinet Office, summary of the seventh meeting of the Expert Panel (held 1 December 2025), in Japanese. https://www8.cao.go.jp/cstp/kokusaiteki/integrity/yushikisha/7kai/gijirokugaiyo.pdf . Responses on publication of the manual and briefings, English translation, the wording on breaches and literacy education, and specialist staff, training materials and running training. Checked against a copy retrieved on 22 November 2026. ↩ ↩2 ↩3 ↩4
-
Acquisition, Technology and Logistics Agency (ATLA), "FY2026 Innovative Science and Technology Initiative for Security (Commissioned Research) Call for Proposals", March 2026, in Japanese. https://www.mod.go.jp/atla/funding/koubo/r08/r08koubo_full_itaku.pdf . Cover (call period), 2.6.3(2) (compliance and research ethics education), 5.13.3 (questionnaire due around mid-July), 5.13.6 (measures for breaches of the manual), and Annex 5 "Questionnaire on Research Security", items 3(1) (fields 1 to 13), 3(2), 4(1) and 4(2). Checked against a copy retrieved on 20 November 2026. ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10
-
Cabinet Office, summary of the fifth meeting of the Expert Panel (held 18 July 2025), in Japanese. https://www8.cao.go.jp/cstp/kokusaiteki/integrity/yushikisha/5kai/gijirokugaiyo.pdf . A member's point on the scope of routine information collection and the response on a two-stage approach starting with literacy. Checked against a copy retrieved on 22 November 2026. ↩
-
MEXT, Science and Technology Policy Bureau, "Specific Directions for Measures under MEXT-Related Policies to Ensure Research Security at Universities and Other Institutions" (our translation), 18 December 2024, in Japanese. https://www.mext.go.jp/content/20250423-mxt_kagkoku-000039402_1.pdf . Overview sections 2 (no discrimination) and 3(2) (training materials and programs); main text section 4 (basic approach) and 5(2). Checked against a copy retrieved on 22 November 2026. ↩ ↩2
-
Japan Science and Technology Agency (JST), "FY2026 Strategic Basic Research Programs (CREST, PRESTO, ACT-X) Call for Proposals", 7 April 2026 (updated 17 April), in Japanese. https://www.jst.go.jp/kisoken/boshuu/teian/koubo/2026youkou.pdf . Cover (call periods), 6.2 (research ethics education) and 6.5 (designation as a Specified Research and Development Program and the questionnaire). Checked against a copy retrieved on 20 November 2026. ↩ ↩2
-
Cabinet Office, "Measures for ensuring research security" (in Japanese). https://www8.cao.go.jp/cstp/kokusaiteki/effort/security.html . The list of documents posted. Checked on 22 November 2026; no document published as training materials was listed. ↩
-
Cabinet Office, "Research Integrity" (English page). https://www8.cao.go.jp/cstp/english/about/research_integrity.html . Provisional translations of the research integrity checklist templates (for researchers, 17 December 2021 version; for universities and research institutes, 29 June 2023 version). Checked on 22 November 2026; no English version of the manual was listed. ↩ ↩2
-
METI, "Q&A on Clarifying the Control of Deemed Exports" (our translation), revised 8 August 2023, in Japanese. https://www.meti.go.jp/policy/anpo/law_document/minashi/minashiqa3.pdf . Question 59. Checked against a copy retrieved on 20 November 2026. ↩
-
FY2024 MEXT commissioned project on support for building systems to ensure research integrity (near-miss case survey), "Research Integrity Near-Miss Casebook" (our translation), March 2025, in Japanese. https://www.mext.go.jp/content/20250331-mxt_kagkoku-000019002_1.pdf . Researcher case 5, "Unintended co-authorship". Checked against a copy retrieved on 22 November 2026. ↩
-
MEXT, "Research integrity and research security" page (in Japanese). https://www.mext.go.jp/a_menu/kagaku/integrity/index.html . The versions of the research integrity checklist templates. Also the notice "Setting up MEXT's Research Security Consultation Desk" (our translation; MEXT notice No. 54, 10 April 2025), https://www.mext.go.jp/content/20250424-mxt_kagkoku-000039402_1.pdf , sections 1 (who may consult) and 2 (points to note). Both checked against copies retrieved on 22 November 2026. ↩ ↩2






