> For the complete documentation index, see [llms.txt](https://doc.dastra.eu/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://doc.dastra.eu/en/gdpr/concepts/dpia.md).

# Data protection impact assessment (DPIA)

Learn what a data protection impact assessment (DPIA) is.

## Definition

The **data protection impact assessment (DPIA)**, provided for by [Article 35 of the GDPR](https://eur-lex.europa.eu/eli/reg/2016/679/oj#d1e3546-1-1), is a **compliance mechanism** that aims to:

* **Identify and minimize the risks** of infringement on the rights and freedoms of individuals,
* **Prove compliance** with the GDPR,
* And **integrate privacy from the design phase** (*Privacy by design*).

> The DPIA focuses on **risks to individuals**, not risks to the organization.

You may also meet the term privacy impact assessment (PIA) for the same approach.

***

## The three components of a DPIA

A DPIA is structured around three major steps:

1. **Description of the processing** → Objectives, context, stakeholders, technologies, data involved.
2. **Assessment of necessity and proportionality** → Legal compliance analysis: purposes, legal bases, individual rights, retention periods, etc.
3. **Risk study and security measures** → Identification of privacy risks and determination of control measures (technical and organizational).

> The DPIA is simultaneously **legal**, **technical** and **organizational**: it engages the whole project team.

***

## Objective of the DPIA

The DPIA enables an organization to:

* **Demonstrate compliance** (accountability principle),
* **Anticipate risks** and compliance costs,
* **Document choices** and design decisions,
* **Build trust** with users, clients or employees.

***

## When a DPIA is mandatory

A DPIA is **mandatory** when processing is **likely to result in a high risk** to the rights and freedoms of individuals.

### Examples of processing activities concerned

* Intelligent video surveillance,
* Automated profile scoring,
* Large-scale collection of health data,
* Use of biometrics or AI on personal data.

{% hint style="info" %}
As a general rule, processing that meets at least **2 of the following criteria** (European Data Protection Board, EDPB) calls for a DPIA:

* Evaluation or scoring,
* Automated decision-making with legal or similar effects,
* Systematic monitoring,
* Sensitive data,
* Large scale,
* Cross-referencing of data,
* Vulnerable individuals,
* Innovative use (AI, IoT, big data),
* Blocking of a right or service.

A single criterion can be enough for particularly sensitive processing. Supervisory authorities also publish lists of processing operations that require a DPIA: see the [DPIA guidance, country by country](https://doc.dastra.eu/en/other/dpia-by-country.md).
{% endhint %}

***

## Timing and updates

The DPIA must be carried out **before the implementation of the processing**, ideally from the design phase. It is a concrete application of the **Privacy by Design** principle.

It must also be:

* **Reviewed regularly**,
* **Revised upon any significant change** to the processing or to its level of risk (Article 35, paragraph 11),
* **Linked to your record** in Dastra for continuous monitoring.

<figure><img src="https://2697025545-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F-LvBxs22wUMicv9uWp6C-1972196547%2Fuploads%2Fgit-blob-be577c50a43246c5ab62e8f1f5f37d0519466afb%2Fdpia-iterative-process.svg?alt=media" alt="Iterative DPIA cycle in seven steps: description of the processing, assessment of necessity and proportionality, measures to demonstrate compliance, assessment of risks to rights and freedoms, measures to address the risks, documentation, monitoring and review"><figcaption><p>An iterative process</p></figcaption></figure>

> The DPIA is an **iterative process**: it accompanies the entire lifecycle of the processing activity.

***

## Carrying out a DPIA

### 1. Assess the necessity and proportionality of the processing

Ask yourself the right questions:

* Are the **purposes** determined, explicit and legitimate?
* Is the **legal basis** clear?
* Is data **minimized**, accurate and up to date?
* Are **retention periods** limited?
* Are **data subjects** properly informed?

### 2. Review the protection of individuals' rights

* Transparent information, clear consent, right of access, erasure and objection.
* Contracts with processors.
* Safeguards for international transfers.
* Corrective measures planned in case of a breach.

### 3. Analyze privacy risks

For each **feared event**:

* Identify the **potential impacts** on individuals (privacy breach, discrimination, reputational harm…).
* Estimate the **severity** and **likelihood** of the risk.
* Determine the **existing or planned protection measures**.
* Assess the **residual risk** and, if necessary, propose **additional measures**.

{% hint style="success" %}
Final objective: reach an **acceptable level of residual risk**, documented and justified.
{% endhint %}

***

## Stakeholders to involve

| Stakeholder              | Role in the DPIA                                               |
| ------------------------ | -------------------------------------------------------------- |
| **Data controller**      | Bears responsibility for the DPIA and leads its implementation |
| **DPO**                  | Advises on the DPIA and monitors its performance               |
| **CISO / IT department** | Provides technical expertise and security measures             |
| **Business teams**       | Provide operational details of the processing                  |
| **Processors**           | Communicate necessary information for the assessment           |
| **Data subjects**        | May be consulted or represented in certain cases               |

> The DPIA is a collaborative process: it involves legal, technical and operational teams alike.

***

## DPIA and artificial intelligence

The **AI Act** introduces a specific obligation to **document and assess high-risk AI systems**. Organizations must include:

* The origin of training data,
* Robustness tests and bias assessments,
* Planned human oversight controls,
* Transparency measures.

For AI systems, Dastra offers a dedicated impact assessment in the **AI governance** module, under **Impact analysis**, to cover both the GDPR and the AI Act.

***

## Dastra best practices

* Determine whether a DPIA is needed in the **Impact assessment** section of the processing activity, which evaluates the 9 EDPB criteria,
* Start the DPIA from a **questionnaire template** of the library,
* Export the DPIA report as a **PDF** for audits,
* Plan the next review with the **regular reviews** of the processing activity.

{% hint style="success" %}
Dastra guides you step by step through the DPIA with an interactive questionnaire template.
{% endhint %}

***

## For more information

{% content-ref url="/pages/-M6E6sWom2H7F\_vNOCJk" %}
[Impact assessment](https://doc.dastra.eu/en/features/ropa/activity/impact-analysis.md)
{% endcontent-ref %}

{% content-ref url="/pages/rNkNzVSuXi6rd9VHOA1p" %}
[PIA / DPIA](https://doc.dastra.eu/en/features/questionnaires/types/dpia.md)
{% endcontent-ref %}


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://doc.dastra.eu/en/gdpr/concepts/dpia.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
