
You are currently browsing EDRi's old website. Our new website is available at https://edri.org


Subscribe to the bi-weekly newsletter about digital civil rights in Europe.
On 21 October the European Parliament Committee on Civil Liberties, Justice and Home Affairs (LIBE) will vote a hugely important dossier: The General Data Protection Regulation.
This very long legislative document is intended to ensure that our rights to privacy and data protection can be effectively asserted in our everyday lives. One of the main purposes of the Regulation is to give citizens greater control over their personal information - maintaining the principles that were developed in the 1995 Data Protection Directive. Recent revelations have shown just how important this is.
Since more than 4,000 amendments have been proposed in the European Parliament, we are now releasing a series of one-pager documents to explain the most important points of the Regulation. These issue sheets explain which amendments are good for data protection and which amendments might do a lot of harm.
Watch this space, we'll be publishing new documents every two working days! We will, of course, also share them with Members of the Committee.
About these Issue Sheets - an introduction by Prof. Douwe Korff
Download pdf
The Draft EU Data Protection Regulation that will replace the main EC Data Protection Directive (Directive 95/46/EC) is in the process of being adopted. The Regulation will have a major impact on the digital environment for citizens, businesses and public bodies. EDRi and its 35 member organisations in 21 European countries are concerned that any weakening of the European data protection rules and principles will undermine the rights and freedoms of European citizens, both within the EU and internationally, due to the EU’s leading role on privacy issues. The ongoing PRISM/Tempora/Etc. scandal shows how crucial high-level data protection rules are for individual citizens, non-governmental organisations and companies.
The Issue Sheets seek to provide Members of the European Parliament in particular with simple clarifications of the main issues; they are kept to one page on each topic.
There is one issue we should stress in this cover note: the wide “national security” exemptions in the main EU treaties. It should be stressed that these are not absolute, and do not grant Member States total exemption from scrutiny in this regard. The EU Charter on Fundamental Rights, which explicitly demands full protection of personal data, cannot be simply ignored.
Ultimately, it is for the European Court of Justice to determine the scope of the exemption. It is already clear that the activities of the United States’ NSA are manifestly not limited to national security as defined in international law, and it would seem that the same applies to the activities of some EU agencies, such as the UK’s GCHQ. Activities by Member States’ “national security agencies” that are not strictly limited to national security as defined in international law are not covered by the EU treaties’ exemptions, and this should be made clear in the Regulation.
Secondly, we urge the European Parliament to stress that any disclosure of personal data that are within the scope of EU data protection law, such as airline passenger data or data on individuals’ financial transactions, to any national security agency is and will remain subject to the EU rules on the processing of personal data: disclosure is expressly mentioned as a form of processing in both the current Data Protection Directive and in the Draft Data Protection Regulation, and is subject to these instruments. Even if the processing of such data by national security agencies after such a disclosure may be outside EU law (provided it is indeed for strictly-defined national security purposes), the disclosure itself is not: it is and must remain subject to the fundamental principles of European data protection law, including purpose-specification and -limitation, transparency, and transborder data flows. Moreover, the activities of EU Member States’s national security agencies’ activities must still always conform, at the very least, to the minimum European standards on State surveillance, adduced by the European Court of Human Rights (as set out in an Annex to a recent EDRi/FREE submission (pdf) on the matter.
Especially in an online environment, data about users is often not directly identifiable. Online tracking companies do not need or want to know the name of an individual as they care about what a person is are and not who the person is – users are being singled out as worthy of being sold a particular product or accepted for a particular service at a particular price... or not. This happens regardless of how the data is stored: creating a 'pseudonym' will not prevent tracking and analyzing of personal data or taking decisions based on these data.
In order to ensure thorough privacy protection, all personal data including data which is not directly identifiable (like a set of mobile phone numbers) must be given equal protection. So-called 'pseudonymous data' should not fall under a separate regime. The mere fact that data (such as your mobile phone number) does not directly identify you should not mean that it is not worthy of the same protection as your name or your address.
A second set of amendments seek to give less protection to data which are “pseudonymous”, which means that a directly identifiable piece of data is replaced by a pseudonym. This lowering of protection includes all types of data that are 'pseudonimised', including data generated by profiling individuals' personality in online social media, for example. Examples of such amendments are:
Consent should be explicit, specific and informed in all circumstances. In practice, this means that consent should always be strictly linked to the processing that the user was informed about and not include other forms of use of personal data. A user must receive sufficient information to be able to understand the consequences before he or she can give their consent to the processing of their data.
In practice, data controllers should not be able to use "pre-ticked boxes" to gain users' consent for the processing of their personal data, nor infer their consent from other actions such as acceptance of general terms and conditions.
Implicit consent offers less protection, as it can be assumed or included in general terms and conditions. Users will lack a genuine choice about how their data will be processed.
Other amendments introduce a new and unclear concept of "broad consent". This new notion could undermine legal certainty for controllers and would create a loophole, where data subjects agree to something they do not fully understand. Amendments
Clearly specifying that the consent should be both explicit and informed would also be a very good addition. See for example amendment 854 (S&D).
Consent should not only be limited to specific purposes, but also be automatically declared invalid when given in a general and abstract way to unspecified and unpredictable forms of data processing. This is proposed in
It seems reasonably clear that, if businesses or other organisations wish to use personal data, any such use should be based on consent, necessity or a legal obligation. However, there may be cases data processing is appropriate, but does not correspond to these criteria. In these cases, the 'legitimate interest' of the organisation that is processing the data may serve as a basis for data processing. However, such cases should be an exception rather than a rule. We therefore need such exceptional cases to be clearly defined and limited to circumstances that are predictable, well-regulated and sufficiently narrow. By doing so, “legitimate interest” will not become a “loophole” that undermines the credibility of the legislation and the trust of citizens. In particular, the legislation needs to prohibit data from being used for purposes that are not related to the original purpose of collection.
Data minimisation is an essential principle of data protection. It establishes that data collected and processed should not be retained or further used unless this is necessary for purposes that were clearly stated in advance. Before starting to process personal data, controllers should ask themselves: "Do we need this? What for? And for how long?" The existing 95 Directive already contains the principle of "not excessive data collection", but it is not very well observed. This principle is being clarified and strengthened by Article 5 (c) of the proposed General Data Protection Regulation.
When it comes to personal and intimate information held by companies about our private lives, it is essential that these data are kept to a minimum; otherwise we risk more abuse and increase the loss of control over how our personal data are used, shared and sold.
It is important to note that this proposal does not create open-ended rights to have newspaper articles or blogs deleted or to overturn legal obligations on companies to store certain data. There is a specific exception in Article 80 on freedom of speech, and this article can be strengthened and clarified to minimise any risk of misunderstanding. For clarity, the wording “erasure” should be used throughout, as “the right to be forgotten” has led to a great deal of confusion.
Suggested exemptions for anonymous and pseudonymous data - such as amendment 1420 by the EPP - do not make sense; this could increase the liability of intermediaries and wrongly incentivise them to monitor and delete information for which they are not responsible. Furthermore, the drafters appear not to understand that anonymisation, by definition, cannot be undone.
Some amendments have been tabled in order to delete Article 17, paragraph 2 since this provision, as proposed by the Commission, is unclear and could increase the liability of intermediaries and wrongly incentivise them to monitor, restrict or delete information over which they have no control.
This right will make it easier for users to switch services, effectively preventing “lock-in” effects which are harmful for consumers and competition. For instance, if a user is dissatisfied with the service offered by a social network, they would be free to move their data to a different platform, instead of being forced to delete their profile and start all over again.
Data portability would help to stimulate competition by making market entry easier for new companies, as consumers will be more willing to try out these services if this is made easier than it is at the moment. It would also create opportunities for innovative new services, such as a service that would analyse a user's electricity usage to work out if another company would be cheaper or to determine how to consume energy more efficiently.
The formats in which data are provided should be interoperable. Otherwise the right to data portability would make little sense because data could not easily be transferred from service to service.
Finally, it should be clarified that data controllers should not store data that are no longer needed only in order to be able to comply with a possible future request to move the data.
A set of amendments propose vague exceptions that would allow controllers to circumvent this obligation. These include amendments 1497 (ALDE), 1498 (EPP) and 1505 (ALDE)
Requiring the data subject to pay a fee to be able to exercise data portability would create obstacles to exercise this right, thereby undermining its use in practice. Amendment 1500 (EPP) suggests this restriction.
Data portability should not be confused with interference with intellectual property rights or trade secrets such as in amendments 1504 (ECR) and 1512 (EPP).
The Regulation should not omit the public sector from this obligation as proposed by Amendment 1522 (EPP).
The Rapporteur also proposed in amendment 143 to clarify that data portability does not interfere with the obligation of deletion.
As a result, it is important that profiling be prohibited both online and offline unless certain strict conditions are met, including the consent of the individual. Strong safeguards should be put in place, including the right to be provided with meaningful information about the logic behind the profiling.
It is absolutely crucial for both privacy and that the collection and to avoid a profound chilling effect for online communication that profiling can only be done with consent and with a right of erasure.
Footnotes: (1) http://www.cam.ac.uk/research/news/digital-records-could-expose-intima... (2) See https://www.aclu.org/national-security/european-officials-declare-us-f..., for example
Weak controls on export of data means a weakening of protection of data and an incentive to process data abroad to circumvent EU rules.
Export of data should only be possible when verifiable safeguards are in place that data will be processed in line with the minimum EU standards or better.
Some amendments propose weakening the new EU rules by automatically recognising the applicability of past adequacy decisions even though the rules would have changed (e.g. Amendment 2384 (ALDE)).
Some amendments propose weakening the standards when it comes to processing for historical, statistical and scientific research (for example amendments 3075 (EPP), 3077 (ALDE) and 3094 (ALDE)).
Some amendments seek to recognise an adequate protection when the controller has “adduced appropriate safeguards”, and give data controllers the right to unilaterally overrule a decision of non-adequacy by the European Commission (such as amendment 2145 (ALDE) and 2148 (ALDE).
The review of current adequacy decisions is also a good addition, as proposed by amendments 2412 (EPP) and 250 (Greens). Amendment 259 from the Greens seeks to reintroduce the “Article 42” provision that would explicitly prohibit the transfer of personal data to third country law enforcement authorities outside of agreed legal frameworks.
The prohibition of transfer to third countries when the laws allow processing that would be unlawful under the Regulation, as proposed in amendments 2385 (GUE/NGL) and 2386 (S&D) is also welcome.
The Safe Harbour agreement is so unclear that the European Commission is certain that only Recital 90 of the draft Regulation was needed in order to clarify the situation. Meanwhile, the Irish data protection authority is clear that companies are completely within their rights to transfer European data to the US authorities for apparently any purpose, without information being provided to the citizen and regardless of the presence or absence of any safeguards in the United States.(2) This approach profoundly and fatally undermines the fundamental right of European citizens to protection of their personal data.
Footnotes: (1) See https://www.privacyinternational.org/global-data-protection-map (2) http://www.europe-v-facebook.org/Response_23_7_2013.pdf
The provisions must be further clarified, in order to indicate that data protection by design and default relate to both (a) technical measures relating to the design and architecture of the product or service and (b) organisational measures, which relate to operational policies of the controller.
Other amendments describe these principles as a burden for companies, despite the fact that this approach will help to avoid situations in which data protection requirements are an afterthought to the development process, which can result in both higher development costs for companies and lower protection for the data subject. Data protection by design can be more accurately seen as an investment that greatly reduces the privacy risks for both companies and citizens (e.g. costly data breaches).
Moreover, some of the suggested additions are vague and undermine legal clarity and opt for a more self-regulatory approach, which is not sufficient to ensure the implementation of strategies that protect privacy by design and by default.
Examples of such amendments include:
These amendments are:
The level of sanctions is very low in most European countries, and therefore even large companies, whose business is the collection and processing of data, have little incentive to respect the legislation. The European Union needs an effective, uniform and predictable level of enforcement.
In an era of "big data", with all the risks that this implies for the fundamental rights of citizens, regulators and courts must have the power, when necessary, to impose appropriate levels of penalties. The establishment of comprehensive and streamlined remedies is an essential element of the Regulation.
This incident shows quite clearly that the Commission proposals to introduce mandatory notifications to data protection authorities and to the victims of the breach should be supported. Existing laws in the US show that mandatory breach notifications are an effective tool to force companies and other organisations to quickly and comprehensively address breaches, as well as acting as an incentive for better security practices.
Moreover, supervisory authorities should maintain a public register of breaches. The safeguards against excessive notifications to victims and against excessive demands (companies are only required to act “where feasible”) being made of the data controller are clear and reasonable. Any weakening appears entirely unnecessary.
Other amendments limit the breach notification to a narrow set of circumstances and suggest that processors should decide whether the breach needs to be notified or not to the data protection authority or the victim of the breach. Some amendments even go as far to delete the obligation to keep records of breaches. Amendments:
Amendments: