Skip to main content
Version: 3.0.0

Data usage license

License terms​

The AI-READI dataset is shared under a custom license that has been developed as part of the project. The full license terms are available here. It is recommended to read them first to ensure that your envisioned use of the AI-READI dataset fits within the requirements of the license.

Overview​

A brief summary of what you can and can't do with this version of the dataset are provided below.

Can:

  • Use the dataset for type 2 diabetes related research
  • Use the dataset for training AI/ML models and other computational models

Can't:

  • Use the dataset for research unrelated to type 2 diabetes
  • Attempt to identify study participants

FAQ​

Which Agreement to Use​

  1. Why does AI-READI have two different data agreements that look so similar?

    While the Data Transfer and Use Agreement (DTUA) and the Data License Agreement (DLA) appear similar and have the same core purpose, they are used for different data sets that need to be treated differently for legal reasons.

  2. When will I use the Data Transfer and Use Agreement (DTUA)?

    You use the DTUA in two circumstances:

    a) when you wish to download ”restricted” data. “Restricted” data is sensitive de- identified human subjects data including genomic, race and ethnicity and other types of information determined by AI-READI to require special handling.

    b) when you wish to use data for a purpose other than diabetes research.

  3. Do I need approval to use restricted data?

    Yes, both of the potential data uses described in Question 2 above require approval of the AI-READI Data Access Committee (DAC), which will review your specific research proposal and inform you of any inquiries for additional information.

  4. When will I use the Data License Agreement (DLA)?

    The DLA is a “browsewrap” license agreement that you enter into (see below) when you wish to use AI-READI data that is not “restricted data” (see Question 2, above).

  5. Do I need approval to use unrestricted data?

    No. If you wish to download unrestricted data (i.e., not described in Question 2 above), you may do so provided that you comply with the terms of the DLA (see below).

The Data License Agreement (DLA)​

  1. Is the DLA a legally binding contract?

    Yes. By entering your name as the Licensee and clicking “I ACCEPT”, you agree to the DLA and are bound by its terms.

  2. Do I need to sign the DLA?

    You do not need to “sign” the DLA, only click “I ACCEPT” after completing all required information in the AI-READI Data Request Workflow.

  3. Can members of my lab or team use or access the data?

    Yes, but only if you have identified an officially-recognized group, team, subdivision or business unit of your Institute/Employer in the form. This is called an Authorized Group. If you have identified an Authorized Group, you may authorize official members of that Authorized Group (“Authorized Users”) to access and use the Data to the same extent that you have been so authorized. Authorized Users may access and use the Data only so long as they remain Authorized Users. Once they leave the Authorized Group, their access and use privileges automatically terminate.

  4. Who is responsible if a Group Member violates the terms of this Agreement?

    You are responsible and legally liable for the access and use of the Data by all Authorized Users, including any continued use after an Authorized Users leaves the Authorized Group.

  5. Does the DLA legally bind my university or employer?

    No. The DLA only binds you and the members of the Authorized Group that you indicate in the form. You warrant that you have the authority to enter into the DLA and to bind the members of your Authorized Group to comply with it.

  6. Are the terms of the DLA negotiable?

    No. If you do not agree with any of the terms of the DLA, then you should not download or use the data.

  7. Where can I find a copy of the DLA?

    The DLA is available here: https://zenodo.org/records/10642459

The Data Transfer and Use Agreement (DTUA)​

  1. Is the DTUA a legally binding contract?

    Yes.

  2. Do I need to sign the DTUA?

    An authorized representative of your institution needs to sign the DTUA. If that isn’t you, then you should ask a Department Chair or someone in your technology transfer office or legal office who can sign on behalf of the institution.

  3. Who is bound by the DTUA?

    The DTUA binds your institution and, by extension, all employees (faculty, staff, students) of the institution.

  4. What is the significance of the DTUA Attachments?

    The base DTUA is a standardized document released by the National Institutes of Health and is the same across all NIH-funded data projects of this type. Attachment 1 to the DTUA is a description of the Data that you wish to access and the specific project for which that access has been approved by the AI-READI DAC. Attachment 2 to the DTUA imposes the same conditions on uses of data provided under the DTUA as the DLA (see below). Attachment 3 to the DTUA identifies any external collaborators who are permitted to access and use the data under the DTUA.

  5. Are the terms of the DTUA negotiable?

    The standardized terms of the DTUA are uniform across all DTUAs and cannot be altered. However, specific information regarding your approved uses, the duration of your agreement and the nature of the data will be inserted into the DTUA based on the information that you provide to the DAC and the DAC’s determinations.

  6. Where can I find a copy of the DTUA?

    A template DTUA (without Attachments) is available for review here: https://thefdp.org/wp-content/uploads/dtua_March-2024-v05.03.2024-final.pdf. The actual DTUA that you will sign, with all Attachments, will be prepared by AI-READI and provided to you following DAC approval of your request.

Using the Data​

  1. What types of projects may I use the data with?

    Data covered by a DLA may be used for any commercial or non-commercial research purpose that does not violate the specific restrictions of the DLA. Data covered by a DTUA may be used only for the Project specified in the applicable Attachment 1.

  2. On what equipment/servers may I download, use and store the data?

    Both the DLA and the DTUA permit use and storage of the data only on (a) servers and devices maintained by and located within your institution or (b) on cloud or remote storage and backup services (e.g., Dropbox, Google Drive, AWS, Microsoft Azure) operated by a company that has a HIPAA-approved Business Associate Agreement in place with your institution.

  3. Can I download/store the data on my personal laptop or device?

    In the case of data covered by a DLA, you may download, use and store the data on computer equipment that is owned or controlled by the person or entity that has entered into the DLA. In the case of data covered by a DTUA, you may download, use and store the data on computer equipment that is owned or controlled by the entity that has entered into the DTUA.

  4. Can I share the data with someone in my lab?

    Yes, if that person is an employee (faculty, staff, student) of the institution that is bound by the DLA or DTUA, as applicable.

  5. Can I share the data with a visiting scholar in my lab?

    In the case of data covered by a DLA, no. The visiting scholar, or their institution, should enter into a DLA to access the data directly. In the case of data covered by a DTUA, Attachment 3 may permit sharing with specified Collaborator Personnel. If the visiting scholar does not fit within the definition of Collaborator Personnel, then data may not be shared with them, and the visiting scholar’s institution should enter into its own DTUA to gain access to the data.

  6. May I share the data with a collaborator at another institution?

    In the case of data covered by a DLA, no. The collaborator, or their institution, should enter into a DLA to access the data directly. In the case of data covered by a DTUA, Attachment 3 may permit sharing with specified Collaborator Personnel. If the collaborator does not fit within the definition of Collaborator Personnel, then data may not be shared with them, and the collaborator’s institution should enter into its own DTUA to gain access to the data.

  7. Can I use a LLM to analyze the data?

    There is no restriction on the types of tools that can be used to analyze the data. However, no restricted or unrestricted data may be uploaded to a third party LLM vendor or operator (e.g., OpenAI) unless the third party has a HIPAA-approved Business Associate Agreement in place with your institution.

  8. Can I use the data to train a LLM?

    No. The data may not be used for general LLM training purposes nor uploaded to a third party LLM vendor or operator (e.g., OpenAI) for training purposes. If you wish to fine tune an in-house instance of a LLM using the data, this is permitted. You must also take all reasonable precautions to prevent the LLM from memorizing the data.

  9. Can I use the data for a course I am teaching?

    If you wish to permit students to access the data that you have downloaded, then those students must be legally bound by your institution’s policies including the DLA or DTUA, as applicable, and the students must agree to all applicable restrictions on use of the data. If you wish students to download data covered by a DLA themselves from AI-READI, then those students must enter into a DLA individually.

  10. Are open source or open weight models covered differently under the license?

    No. The above restrictions apply to all models.

  11. Do these restrictions apply to derivative data?

    Yes, all restrictions in the DLA and DTUA apply with equal force to derivative data. “Derivative data” is AI-READI data that has been modified, excerpted, encrypted, condensed, encoded, translated or otherwise altered, such that it contains original data or original data may be derived from it. Synthetic data–artificially generated data that mimics real-world data characteristics but doesn't contain any actual personal information–that uses AI-READI data or derivative data is also considered derivative data.

  12. Can I work on privacy issues using the dataset if it involves attempting to re- identify individuals?

No, you are prohibited from attempting to re-identify individual data subjects.

Publication​

  1. May I publish the results of studies that I conduct using the data?

    Yes, so long as you do not replicate more than a small representative sample of the data in your publication or any supplement, online or otherwise.

  2. How should I acknowledge use of the data in papers and presentations that were developed using the data?

    All papers and presentations that were developed using the data must acknowledge the data source as indicated on the data file itself. In addition, you should follow the data citation guidance that can be found here: https://docs.aireadi.org.

Intellectual property​

  1. Do I own the data that I download or that is otherwise provided to me by AI- READI?

    No. Ownership of all data, including any derivative data, remains with us or our suppliers.

  2. Can I acquire intellectual property rights (e.g., patents) in discoveries or inventions that I make using the data?

    Yes. We claim no rights in your discoveries or inventions made using the data, so long as you do not seek to claim any rights in the data itself or any derivative data.

Liability​

  1. What happens if the data contains errors or omissions?

    We make the data available for free, and you use the data entirely at your own risk. We make no warranties regarding the accuracy or completeness of the data.

  2. What happens if I inadvertently cause someone harm by using the data (e.g., if it leaks or leads to re-identification of an individual)?

    You are fully responsible and liable for your use and handling of the data. In addition, if you cause someone harm through your use or handling of the data and we are sued or otherwise liable as a result, you are required to indemnify us against that harm (i.e., you must cover our costs and expenses and any damages or fines assessed against us).