DPDP Rules 2025Updated 22 November 202611 min read

DPDP Exemption for Research, Archiving and Statistics (Rule 16)

Rule 16 switches off the DPDP Act for processing necessary for research, archiving or statistics, but only if you follow the Second Schedule standards. Here is what that means in practice.

By Suhas Bhairav · Reviewed against the DPDP Rules, 2025 (G.S.R. 846(E), 13 Nov 2025)

Provision
Rule 16, DPDP Rules 2025
Parent section
Section 17(2)(b), DPDP Act 2023
Condition
Processing must follow Second Schedule standards
Covered purposes
Research, archiving or statistical purposes
In force from
13 May 2027
Effect
Provisions of the Act do not apply to qualifying processing

Quick answer

Under Rule 16 of the DPDP Rules, 2025, the DPDP Act does not apply to processing of personal data necessary for research, archiving or statistical purposes, provided it is carried on in accordance with the standards in the Second Schedule. The exemption is conditional, not automatic, and Rule 16 comes into force on 13 May 2027 along with the core fiduciary obligations.

  • Rule 16 exempts research, archiving and statistical processing from the whole Act, but only when the Second Schedule standards are met.
  • The Second Schedule requires lawful processing, purpose limitation, data minimisation, accuracy, limited retention, security safeguards and accountability.
  • The exemption is anchored in section 17(2)(b) of the Act, which also bars using such data for decisions specific to an individual.
  • Rule 16 commences on 13 May 2027, eighteen months after the Rules were published on 13 November 2025.
  • Commercial analytics dressed up as research is the main risk; document necessity and keep research datasets separated from production.

Who this is for: Academic and clinical research teams, Data science and analytics teams, Archives, libraries and museums, Survey and market research firms, AI teams building evaluation or training datasets, Privacy and compliance leads

What does Rule 16 actually say?

Rule 16 is one sentence long: the provisions of the Act "shall not apply to the processing of personal data necessary for research, archiving or statistical purposes if it is carried on in accordance with the standards specified in Second Schedule."

Three conditions are packed into that sentence:

  1. Purpose — the processing must be for research, archiving or statistical purposes.
  2. Necessity — the processing must be necessary for that purpose, not merely convenient.
  3. Standards — it must be carried on in accordance with the Second Schedule.

If all three hold, the Act's obligations (notice, consent, Data Principal rights, breach intimation and so on) do not apply to that processing. If any one fails, you are back to ordinary rules: you need a lawful ground under section 6 or 7 and must meet every section 8 obligation.

Rule 16 implements section 17(2)(b) of the DPDP Act, which empowers this exemption. The Act's own wording also excludes cases where the personal data is used to take a decision specific to a Data Principal, so the exemption is built for aggregate, population-level work rather than individual profiling.

Rule 16 is one of the rules that comes into force eighteen months after publication, on 13 May 2027 (Rule 1(4)). Until then, the Act's substantive obligations are not yet in force either, so the exemption and the duties it relieves start on the same day.

Which Second Schedule standards must research follow?

The Second Schedule is shared by Rule 5 (State processing for subsidies and benefits) and Rule 16. It requires "appropriate technical and organisational measures" to ensure effective observance of the following, in accordance with applicable law:

ItemStandardWhat it means for a research dataset
(a)Processing is lawfulYou obtained the data lawfully and no other law prohibits the use
(b)Processing is for the purposes in section 17(2)(b)A documented research, archiving or statistical purpose
(c)Limited to necessary personal dataDrop direct identifiers and fields the analysis does not need
(d)Reasonable efforts at completeness, accuracy, consistencyData quality checks and provenance records
(e)Retained only till required, or as law requiresA defined end date for the dataset
(f)Reasonable security safeguardsAccess control, encryption, logging, including at processors
(h)Accountability of whoever determines purpose and meansA named owner who answers for compliance

Item (g), which requires intimation to the Data Principal and a contact person, is expressly limited to processing "under clause (b) of section 7" — the State benefits case. On a plain reading it does not apply to Rule 16 research, though publishing a contact point is still good practice.

Notice that these standards mirror the Act's own principles. The exemption removes the consent and rights machinery, not the duty to handle data carefully.

Who can rely on the research exemption?

Rule 16 does not restrict the exemption to universities, government bodies or non-profits. It is framed by purpose, not by the type of entity. A private company can, in principle, rely on it for genuine research or statistical work.

Typical candidates:

  • Academic and clinical research using patient, student or survey data, subject to sectoral ethics and health-data rules.
  • Official and industry statistics, such as aggregate reports on usage, demographics or market trends.
  • Archiving of records for historical or reference value, for example newspaper archives or institutional records.
  • Model evaluation and benchmarking where outputs are aggregate metrics rather than decisions about a person.

Cases that are unlikely to qualify:

  • A/B tests and analytics whose output changes what a specific user sees, pays or is offered.
  • Credit, fraud or risk scoring of identified individuals.
  • Building marketing segments or lookalike audiences.
  • Training a model whose purpose is to make individual-level decisions in production, unless the research stage is genuinely separated and documented.

The deciding question is whether the processing is necessary for an aggregate research, archiving or statistical purpose and whether its results are used to act on a particular Data Principal. If you need an individual-level decision, use consent or a section 7 legitimate use instead. For AI-specific cases see DPDP and LLM training data and synthetic data under DPDP.

How to engineer a research pipeline that meets the standards

Treat the Second Schedule as a design spec for a research enclave rather than a legal footnote.

Separate the environment

  • Copy data out of production into a dedicated research store with its own access-control list.
  • Strip direct identifiers (name, phone, email, customer ID) and replace them with keyed pseudonyms; keep the key outside the research environment. See pseudonymisation in AI pipelines.
  • Block any write path from the research store back into production user profiles.

Document necessity

  • Maintain a short research protocol per project: question, purpose category (research, archiving or statistics), fields used and why each is needed, expected outputs, end date.
  • Record the named owner under item (h).

Apply the same security baseline

Item (f) requires reasonable security safeguards. Mirror the Rule 6 controls — encryption or tokenisation, access control, access logs with review, backups — described in reasonable security safeguards.

Control retention

  • Tag every dataset with an expiry date and automate deletion or archival when the project ends.
  • For archives, record the legal or institutional basis that justifies long-term retention.

Output review

Release only aggregates. Apply minimum cell sizes or suppression for small groups so that published statistics cannot be traced back to an individual.

Common mistakes with the research exemption

Assuming the exemption is automatic. Rule 16 applies only "if" the Second Schedule standards are followed. A team that cannot show minimisation, security and accountability records cannot claim it after the fact.

Labelling product analytics as research. Dashboards that drive per-user decisions, personalisation and pricing experiments are product processing. They need a ground under section 6 or 7 and must comply with notice, rights and retention duties.

Re-using research outputs against individuals. Once a research dataset is used to flag, score or contact a specific person, the activity moves outside the exemption. Keep a hard boundary between research findings and operational action.

Keeping data forever "for research". Item (e) ties retention to the purpose. An undefined future research use is not a purpose.

Ignoring processors. Item (f) covers processing "undertaken by it or on its behalf by a Data Processor". Cloud notebooks, labelling vendors and survey platforms need contractual security obligations.

Forgetting other laws. The exemption relieves DPDP obligations only. Sectoral rules on health records, clinical trials, ethics committee approvals and official statistics still apply, and item (a) requires the processing to be lawful under them.

Confusing Rule 16 with publicly available data. Scraped public data is handled by a different carve-out in section 3(c); see publicly available data.

How Rule 16 interacts with other DPDP provisions

Rule 16 sits within a wider set of exemptions and obligations:

  • Section 17 exemptions. Rule 16 is the operative standard for section 17(2)(b). Other section 17 exemptions — legal claims, offences, court-approved mergers, notified State instrumentalities — have their own conditions. See section 17 exemptions.
  • Rule 5. The same Second Schedule governs State processing for subsidies, benefits, certificates, licences and permits under section 7(b). Government research units may touch both. See Rule 5 State processing.
  • Rule 8(3) log retention. For processing that is not exempt — for example, the production system the research data was extracted from — the one-year minimum retention of personal data, traffic data and logs still applies. See one-year log retention.
  • Significant Data Fiduciaries. If you are notified as an SDF, your annual DPIA under Rule 13 is a natural place to record which research activities rely on Rule 16 and why.
  • Children's data. Research that touches children's data and fails the standards falls back to section 9, including verifiable parental consent under Rule 10.

Because Rule 16 removes the entire Act, the Board will expect strong evidence when an organisation relies on it. Treat your protocol, minimisation record and access logs as the evidence file.

What to do before 13 May 2027

The exemption and the obligations it relieves both commence on 13 May 2027. Use the remaining time to decide, project by project, which processing will rely on Rule 16 and which will rely on consent or a legitimate use.

A practical sequence:

  1. Inventory every research, analytics, archive and statistics workload. Your data map should have a column for "lawful basis or exemption".
  2. Classify each as qualifying research, archiving or statistics, or as product processing.
  3. Write protocols for the qualifying ones: purpose, fields, necessity, owner, end date.
  4. Build or harden the research enclave: pseudonymisation, access control, logging, no write-back.
  5. Update processor contracts to include security safeguards for research vendors and cloud notebooks.
  6. Plan retention: set expiry dates and automated deletion.
  7. Move non-qualifying analytics under your consent notice or a legitimate use, and make sure your privacy notice covers them.

Teams that complete this before May 2027 will be able to keep research running without bolting consent flows onto aggregate analysis that never needed them.

Further reading on suhasbhairav.com

Before 13 May 2027

Compliance checklist

  1. List all research, archiving and statistics workloads

    Record data sources, fields, owners and outputs for each workload in your data inventory.

    Rule 16

  2. Document the purpose and necessity of each dataset

    Write a short protocol stating the research, archiving or statistical purpose and why each field is needed.

    Second Schedule items (b) and (c)

  3. Minimise and pseudonymise

    Remove direct identifiers and fields that are not needed; keep re-identification keys outside the research store.

    Second Schedule item (c)

  4. Run data quality checks

    Make reasonable efforts to keep research data complete, accurate and consistent, and record provenance.

    Second Schedule item (d)

  5. Set a retention end date

    Tag each dataset with an expiry and automate deletion or archival when the purpose is achieved.

    Second Schedule item (e)

  6. Apply security safeguards, including at processors

    Use encryption or tokenisation, access control and access logs, and put security clauses in vendor contracts.

    Second Schedule item (f)

  7. Name an accountable owner

    Assign the person who determines purpose and means and who answers for compliance with the standards.

    Second Schedule item (h)

  8. Block individual-level use of research outputs

    Prevent research results from being written back to user profiles or used to decide about a specific person.

    Section 17(2)(b)

Key dates

  1. DPDP Act, 2023 receives assent, including the section 17(2)(b) research exemption power

  2. DPDP Rules, 2025 published as G.S.R. 846(E)

  3. Rule 16 and the Second Schedule standards come into force

Answers

Frequently asked questions

Does the DPDP Act apply to research data?

Not if the processing is necessary for research, archiving or statistical purposes and is carried on in accordance with the Second Schedule standards. In that case Rule 16 says the Act's provisions do not apply. If the standards are not met, or the data is used to decide about a specific individual, the Act applies in full and you need consent or a legitimate use.

Can a private company use the DPDP research exemption?

Yes, in principle. Rule 16 is framed around the purpose of processing, not the type of organisation, so a company doing genuine research or statistical work can rely on it. It must still show that the processing is necessary for that purpose and that it meets every applicable Second Schedule standard, including minimisation, security and accountability.

Do I need consent to use customer data for statistics under DPDP?

Not if the statistical processing qualifies under Rule 16, which takes it outside the Act. That requires necessity, aggregate outputs and compliance with the Second Schedule standards. If the statistics feed decisions about specific customers, such as pricing or targeting, the exemption does not fit and you need consent or a section 7 legitimate use.

When does the Rule 16 research exemption start?

Rule 16 comes into force on 13 May 2027, eighteen months after the DPDP Rules were published on 13 November 2025. That is the same date on which the Act's core fiduciary obligations, such as notice, security safeguards and breach intimation, begin to apply.

Is AI model training covered by the DPDP research exemption?

Only where the training is genuinely necessary for a research or statistical purpose and follows the Second Schedule standards. Training a model that will be used to make decisions about individuals in production is harder to fit within the exemption. Document the purpose, separate research and production, and seek legal advice on borderline cases.

Do Data Principal rights apply to exempt research datasets?

No. When Rule 16 applies, the provisions of the Act do not apply to that processing, which includes the rights to access, correction, erasure and nomination. The organisation must still follow the Second Schedule standards, so retention must end when the purpose is achieved and security safeguards must be maintained throughout.

Legal references and sources

  • Digital Personal Data Protection Rules, 2025 — Rule 16 (official text)
  • Digital Personal Data Protection Rules, 2025 — Second Schedule (official text)
  • Digital Personal Data Protection Act, 2023 — Section 17(2)(b) (official text)
  • Digital Personal Data Protection Rules, 2025 — Rule 1(4) (commencement) (official text)

This guide explains the law for engineering and business teams. It is not legal advice; confirm how it applies to your organisation with qualified counsel, and check the Gazette for later notifications or amendments.

About the author

Suhas Bhairav

Suhas Bhairav

Systems architect and applied AI practitioner who designs production AI systems for Indian and global companies, with a focus on privacy-by-design under the DPDP Act and the EU AI Act.