# 12 Data protection impact assessment – a key part of GDPR compliance

Further to the series of publications regarding the changes introduced by the GDPR, in this publication we will introduce you to one of the new concepts set out in the GDPR, namely the Data protection impact assessment (DPIA).

What is DPIA?

Data controllers are responsible for introducing appropriate safeguards to ensure compliance with the GDPR taking into account “the risks of various likelihood and severity to the rights and freedoms of natural persons”. In this sense, their role is not limited solely to the control and definition of the purposes and means of personal data processing, but also includes their obligation to manage the risks that could arise as a result of that activity.

The main objective of the DPIA is to clarify, to describe all processing processes, to assess their necessity and proportionality, and to contribute to the adequate and appropriate management of risks to the rights and freedoms of natural persons arising from the processing of their personal data.

What is DPIA expressed in and what does it contain?

Article 35, Para.7 of the GDPR sets out the minimum features of a DPIA, namely:

  • a systematic description of the envisaged processing operations and the purposes of the processing;
  • an assessment of the necessity and proportionality of the processing operations in relation to the purposes;
  • an assessment of the risks to the rights and freedoms of data subjects;
  • the measures envisaged to address the risks and demonstrate compliance with thе Regulation.

The assessment of the risks to the rights and freedoms of the processing is one of the main components of the DPIA. Some of the risk assessment guidelines and principles set out in the GDPR partially overlap already existing internationally recognised risk management standards, such as ISO 31000:2009. Examples in this regard are:

  • establishing the context: “taking into account the nature, scope, context and purposes of the processing and the sources of the risk”;
  • assessing the risks: “assess the particular likelihood and severity of the high risk”;
  • treating the risks: “mitigating that risk” and “ensuring the protection of personal data”, and “demonstrating compliance with this Regulation”.

DPIA – when it should be carried out?

It is important to note that the carrying out of a DPIA is only mandatory where a processing is “likely to result in a high risk to the rights and freedoms of natural persons”. In this sense, considered to be the most dangerous is the processing of data that is of a very personal nature by technical means without human intervention (e.g. algorithms / software) including:

  •  Systematic and detailed assessment of personal aspects related to health, workplace performance, personal preferences, location, economic status with respect to individuals, which is based on automatic processing, including profiling, for creating or using personal profiles;
  • Large-scale processing of special categories of data (e.g. data on racial or ethnic origin, political views, sex life, etc.) or personal data on convictions and offenses (Article 9, Para. 1 and Article 10 GDPR);
  • A systematic monitoring of a publicly accessible area on a large scale; etc.

In any case, the DPIA should be carried out before the processing. (Article 35, Para.1 and 10, Rec. 90 and 93). The Working Party under Art. 29 recommends that evaluation should be carried out, even when there is doubt as to the need for such assessment, as “DPIA is a useful tool to help controllers comply with data protection law” [1] and with the principle of accountability (for more information see publication #8 ACCOUNTABILITY AS A NEW PRINCIPLE OF GDPR).

Which data processing activities are considered high risk?

In the Guidelines of the Working Party under Art. 29 there are nine criteria in which the controller can identify operations that may result in a high risk to the rights and freedoms of the natural persons:

1) Existence of assessment or scoring, including profiling and prediction – an example in this respect are financial institutions reporting to their clients in connection with granting credits in databases for fighting against money laundering or terrorist financing; genetic testing for prediction of disease risks; companies that create behavioural or marketing profiles based on a website; etc.

2) Existence of automated decision making with legal or similar significant effect – automated decision making is the ability to make decisions by technological means without any human involvement. An example is when the decision on credit approval is made by a person on the basis of a profile designed entirely by automated means, or when the decision on the approval of the loan is made by means of an algorithm and the person is automatically notified of the decision without first being made a meaningful assessment by a human.

3) Existence of processing used for different types of surveillance or control of data subjects – including monitoring through which personal data is being processed where data subjects do not realize who collects their data and how it will be used, e.g. video surveillance in a public area;

4) Existence of processing of special categories of data – public hospitals that store patients’ medical records should carry out a DPIA because they operate with sensitive data, notably the health of natural persons;

5) Existence of large-scale processing of personal data – determining the scale is a separate process that involves careful consideration of factors such as: the number of data subjects involved, the volume of data or the scope of the different types of data, the duration or continuity, and the geographical scope of the processing activity;

6) Datasets that have been matched or combined, resulting from two or more processing operations performed for different purposes and / or by different controllers in a way that goes beyond the reasonable expectations of the subject – In such cases, the nature of the contractual arrangements, and the balance between the subject and the data controller in particular, should be examined, for example, to what extent the data subject is free to terminate the contract and seek alternative service providers;

7) Processing of data concerning vulnerable data subjects including any case where imbalance between the position of the data subject and the controller can be identified – examples in this respect are children – persons subject to treatment, mentally ill persons, asylum seekers, patients, elderly people, etc.

8) Existence of innovative use or applying technological or organisational solutions – an excellent example in this regard is the use of fingerprints and face recognition to improve access control;

9) Preventing data subjects from exercising the right to use a service or contract – here again is the example with a bank screens a client aganst a credit reference database, i.e. in this case, the processing of the subject’s personal data may lead to them being deprived of the possibility of taking a loan.

As a rule of thumb, a processing operation meeting less than two of the aforementioned criteria may not require the carrying out of a DPIA due to the lower level of risk. Conversely, the more criteria the processing operations meet, the higher the likelihood of high risk with regard to the rights and freedoms of natural persons.

In addition, GDPR imputes an obligation to the supervisory authority for establishment and publishing a list of the processing operations that require a DPIA. Moreover, the supervisory authority could establish and publish a list with the processing operations, for which a DPIA is not mandatory. Currently, the Commission for Personal Data Protection (CPDP) has not made such lists public yet.

What follows after the carrying out of DPIA?

DPIA does not constitute of a single action, but of an entire process of achieving and demonstrating compliance.

However, once carried out, the DPIA could be applied for assessment of numerous processing operations similar to the risks presented, taking into account the nature, scope, context and purposes in light of the concrete case. When the processing operation involves joint controllers, they need to define their respective obligations precisely and clearly. Their DPIA should set out which party is responsible for the various measures designed to treat risks and to protect the rights of the data subjects.

If the data controller considers a DPIA not to be mandatory, he is obliged to make a detailed statement of the reasons for not taking action.

Carrying out the DPIA may be outsourced to an outside person.

If the processing is wholly or partly performed by a data processor, the processor should assist the controller in carrying out the DPIA and provide any necessary information. The roles and responsibilities of the processors must be defined precisely in a separate contract/agreement. Such obligations are usually imposed to the processor in the data processing agreement between the controller and the processor.

The controller will also have to consult the supervisory authority whenever Member – state Law requires such actions.

In the light of the above, it remains intriguing what measures would be provided in the new amendments of the Personal Data Protection Act.

In conclusion, it should be emphasized on the fact that carrying out a DPIA is a key part of complying with GDPR in cases of high risk processing. This means that data controllers should be able to determine whether a DPIA has to be carried-out or not. Of course, the internal data controller policy could extend the list of data processing activities for a DPIA will be carried-out even beyond the requirements of the GDPR. Such an approach shall absolutely result in building greater trust and confidence of data subjects in the controller and in providing for additional safeguards for the lawful and adequate processing of personal data within the company.

[1] Data Protection Working Party under Art. 29: Guidelines on Data Protection Impact Assessment (DPIA) and determining whether processing is “likely to pose a high risk” for the purposes of Regulation 2016/679 (WP 248 rev. 01), available here.