Lorna Woods, Professor of Internet Law, University of Essex
This case concerns the collection of bulk communications data (BCD) from network operators by the security and intelligence agencies (SIAs). It formed part of an action brought by Privacy International challenging the SIAs’ acquisition, use, retention, disclosure, storage and deletion of bulk personal datasets (BPDs) and BCD which started in 2015 before the Investigatory Powers Tribunal (IPT). Privacy International’s claim is based on its understanding of the safeguards required by the Court of Justice in Tele2/Watson – a 2016 CJEU judgment on UK data retention law, discussed here.
In Tele2/Watson the Court of Justice held that any data retention obligation must be targeted and limited to what is strictly necessary in terms of the persons affected, the sorts of data retained and the length of retention. It also suggested that access to retained data should be subject to prior review by an independent body and that parties affected should be informed of the processing (unless this would compromise the investigations); and that the data should be retained within the EU. The authorities must take steps to protect against misuse of data and any unlawful access to them. Privacy International argued that the safeguards provided by British law are insufficient. The British government claimed that the SIAs’ activities fell outside the scope of EU law and that the rules were compliant with Article 8 ECHR. It argued that providing the safeguards as required by Tele2/Watson would undermine the ability of the SIAs. The IPT referred two questions – but only in relation to BCD not BPD – to the Court of Justice. This was the basis for the Court’s judgment handed down yesterday.
Questions in Issue
The two questions referred were:
- whether the activities of the SIAs fall within the scope of EU law bearing in mind Art 4 TEU and Art 1(3) of Directive 2002/58 (ePrivacy Directive);
- if the answer is that the situation falls within EU law, do any of the “Watson Requirements” (as above) (or any other requirements) apply?
The Court of Justice decided to deal with this case with two other cases that had been referred to it: Joined cases C-511/18 and C-512/18 La Quadrature du Net & Ors and Case C-520/18 Ordre des barreaux francophones et germanphone & Ors, which were also the subject of a separate judgment yesterday. The cases also dealt with the bulk collection of communications data but in addition the court in La Quadrature du Net also asked whether real-time measures for the collection of the traffic and location data of specified individuals, which, whilst affecting the rights and obligations of the providers of an electronic communications service, do not however require them to comply with a specific obligation to retain their data are permissible. It also asked whether the Charter required persons concerned by surveillance to be informed once such information is no longer liable to jeopardise the investigations being undertaken by the competent authorities, or may other existing procedural guarantees which ensure that there is a right to a remedy suffice? Ordre des barreaux francophones et germanphone & Ors raised the question of whether a general obligation might be justified to identify perpetrators of secual abuse of minors. If national law has not usfficiently guaranteed human rights may the effects of that law be temporarily retained in the interests of certainty and to achieve the objectives set down in the law.
- the e-privacy directive (and EU law in general) applies in this situation because of the required co-operation of private parties;
- limitations on the obligation to guarantee the confidentiality of communications must be interpreted narrowly and with regard to the rights in the EU Charter on Fundamental Rights;
- the case law in Tele2/Watson (summarised above) should be upheld: general and indiscriminate retention of traffic and location data of all subscribers is an interference with the fundamental rights enshrined in the Charter but real-time collection of traffic and location data of individuals suspected of being connected to a specific terrorist threat could be permissible provided it down not impose a requirement on communications service providers to retain additional data beyond that which is required for billing/marketing purposes; and that the use of such data for purposes less serious than the fight against terrorism and serious crime was incompatible with EU law.
Summary of Judgment
In its Grand Chamber judgment, the Court confirmed that requirements on communications service providers to retain data fell within the scope of EU law and specifically the e-Privacy Directive. The Court argued that the exclusion in Article 1(3) e-Privacy Directive related to “activities of the State or of State authorities and are unrelated to fields in which individuals are active” (para 35, citing Case C-207/16 Ministerio Fiscal, discussed here, para 32), whereas Art 3 makes clear that it regulates the activities of communications service providers. As held in Ministerio Fiscal, the scope of that directive extends not only to a legislative measure that requires providers of electronic communications services to retain traffic data and location data, but also to a legislative measure requiring them to grant the competent national authorities access to that data.
The legislative measures, permissible as a derogation under Article 15, “necessarily involve the processing, by those providers, of the data and cannot, to the extent that they regulate the activities of those providers, be regarded as activities characteristic of States” (para 39). given the breadth of the meaning of ‘processing’ under the GDPR, the directions made under s 94 Telecommunications Act fall within the scope of the ePrivacy Directive. The Court re-affirmed (para 43) the approach of its Advocate General in this case (and in La Quadrature du Net) that ‘activities’ in the sense of Art 1(3) cannot be interpreted as covering legislative measures under the derogation provision; to hold otherwise would deprive article 15 of any effect (following reasoning in Tele2/Watson) and Article 4(2) TEU does not disturb that conclusion (despite the Court’s reasoning in the first PNR case (Cases C-317/04 and C-318/04, paras 56 to 59). For the e-Privacy Directive (by contrast to the former Data Protection Directive in issue in the PNR case), what is important is who does the processing; it is the communications providers. The Court took the opportunity to confirm that the GDPR should not be interpreted the same way as the Data Protection Directive but in parallel with the e-Privacy Directive.
As regards the second question, the Court re-stated the scope of s. 94 orders thus:
That data includes traffic data and location data, as well as information relating to the services used, pursuant to section 21(4) and (6) of the RIPA. That provision covers, inter alia, the data necessary to (i) identify the source and destination of a communication, (ii) determine the date, time, length and type of communication, (iii) identify the hardware used, and (iv) locate the terminal equipment and the communications. That data includes, inter alia, the name and address of the user, the telephone number of the person making the call and the number called by that person, the IP addresses of the source and addressee of the communication and the addresses of the websites visited.
Such a disclosure of data by transmission concerns all users of means of electronic communication, without its being specified whether that transmission must take place in real-time or subsequently. Once transmitted, that data is, according to the information set out in the request for a preliminary ruling, retained by the security and intelligence agencies and remains available to those agencies for the purposes of their activities, as with the other databases maintained by those agencies. In particular, the data thus acquired, which is subject to bulk automated processing and analysis, may be cross-checked with other databases containing different categories of bulk personal data or be disclosed outside those agencies and to third countries. Lastly, those operations do not require prior authorisation from a court or independent administrative authority and do not involve notifying the persons concerned in any way.Paras 51-52
The Court stated that the purpose of the e-Privacy Directive was to protect users from threats to their privacy arising from new technologies. It ‘gave concrete expression to the rights enshrined in Articles 7 and 8 of the Charter’ (para 57) and the exceptions thereto under Article 15(1), ie necessary, appropriate and proportionate in the interests of purposes listed in Art 15(1): national security, defence and public security, and the prevention, investigation, detection and prosecution of criminal offences or of unauthorised use of the electronic communication system. The exceptions cannot permit this exception to become the rule (citing Tele2/Watson, but also the ruling in La Quadrature du Net). Restrictions must also comply with the Charter. This is the same whether the legislation requires retention of the transmission of data to third parties (citing EU-Canada PNR Agreement, discussed here, paras 122-123). Drawing on Schrems II, discussed here, the Court held:
any limitation on the exercise of fundamental rights must be provided for by law implies that the legal basis which permits the interference with those rights must itself define the scope of the limitation on the exercise of the right concerned.Para 65
It also re-iterated that derogations from the protection of personal data any restriction on confidentiality of communications and traffic data may apply only in so far as is strictly necessary and “by properly balancing the objective of general interest against the rights at issue’ (para 67). Proportionality also requires the legislation to lay down clear and precise rules governing the scope and application of the measure in question and imposing minimum safeguards, to protect effectively against the risk of abuse. The legislation must set down conditions for the application of the measures so as to restrict them to those ‘strictly necessary’; the legislation must be binding. Automated processing gives rise to greater risks. These considerations are the more pressing in the context of sensitive data.
The Court noted that the transmission of data to SIAs constituted a breach of confidentiality in a general and indiscriminate way and thus:
has the effect of making the exception to the obligation of principle to ensure the confidentiality of data the rule, whereas the system established by Directive 2002/58 requires that that exception remain an exception.Para 69
it also constitutes an interference with Articles 7 and 8 of the Charter, no matter how the data are subsequently used. Re-iterating its approach in EU-Canada PNR Opinion, the Court stated that:
it does not matter whether the information in question relating to persons’ private lives is sensitive or whether the persons concerned have been inconvenienced in any way on account of that interference.Para 70
Here, given the potential to create a personal profile of individuals the intrusions was particularly serious and “no less sensitive than the actual content of communications” (para 71). The court also emphasised the impact of the feeling of being under constant surveillance, following its reasoning in Digital Rights Ireland (discussed here) and Tele2/Watson. Such surveillance may have an impact on freedom of expression, especially where users are subject to professional secrecy rules or are whistleblowers. The Court also note that given the quantity of data in issue, their “mere retention” entails a risk of abuse and unlawful access (para 73).
The Court distinguished between ‘national security’ understood in the light of Article 4(2) TEU and ‘public security’ and matters within Article 15 ePrivacy Directive. While measures safeguarding national security must still comply with Art 52(1) of the Charter, given the seriousness of threats comprised in ‘national security’ in principle the objective of safeguarding national security is capable of justifying more intrusive measures that those would could be justified by other objectives (cross referring to its reasoning in La Quadrature du Net).
Even in relation to national security, the underlying national legislation must also lay down the substantive and procedural conditions governing use of the data and not just provide for access. National legislation must rely on objective criteria in order to define the circumstances and conditions under which the competent national authorities are to be granted access to the data at issue. Here, the national legislation requiring providers of electronic communications services to disclose traffic data and location data to the security and intelligence agencies by means of general and indiscriminate transmission exceeds the limits of what is strictly necessary and cannot be considered to be justified, within a democratic society even in the interests of protecting national security.
La Quadrature du Net/Ordre des barreaux francophones et germanophone
The Court’s approach to Article 15 and the sorts of activities in the service of which surveillance may be undertaken by contrast with Article 3(1) was, unsurprisingly, the same as can be seen in Privacy International, as was its approach to interpreting the directive – emphasising the confidentiality of communications as well as Articles 7 and 8 EU Charter. Again, the Court took the approach that the exception to communications confidentiality should not become the rule and that exceptions must be strictly necessary and proportionate to their objectives. Retention of communications data is a serious interference with fundamental rights – including freedom of expression. The retention of the data constitutes such an interference whether or not the data are sensitive or whether the user was inconvenienced.
In similar terms to Privacy International, the Court again came to the conclusion that the general and indiscriminate retention of data was impermissible under the Charter and Article 15. The Court also re-stated the limitations on derogating measures made under Art 15. The point of difference in this analysis is that the Court recognised the conflicting rights that might need to be reconciled – particularly with regard to crimes against minors and the State’s positive obligation to protect them. This does not mean that the limits as regards necessity and proportionality may be overlooked.
The Court then considered the meaning of national security – approaching the matter in the same terms as it did in Privacy International. This higher threshold meant that neither the directive nor the Charter precludes recourse to an order requiring providers of electronic communications services to retain, generally and indiscriminately, traffic data and location data. This however is only so when the Member State concerned is facing a sufficiently serious threat to national security (which includes matters more serious than those listed in Art 15), a threat that is genuine and actual or foreseeable. In such a case retention can only be for a period of time limited to that which is strictly necessary. If any such order is to be renewed it must be for a specified length of time. The retained data must be protected by strict safeguards against the risk of abuse. The decision must be subject to effective review by an independent body (court or administrative), whose decision is binding, in order to verify that such a situation exists and that the conditions and safeguards laid down are observed.
The Court observed that general and indiscriminate surveillance refers to that which covers virtually all the population. The Court recognised the duties of the State under positive obligations and the need to balance potentially conflicting rights. It then held that in situations such as those described at paras 135-6 of its judgment, that is those falling in Article 4(2) TEU, the e-Privacy Directive and the Charter do not preclude measures for targeted retention of traffic and location data. Such measures must be limited in time to what is strictly necessary, and focused on categories of persons identified on the basis of objective and non-discriminatory factors, or by using geographical criteria. It then relied on similar reasoning in relation to the fight against crime and the protection of public safety.
Similarly, IP addresses may be retained in a general and indiscriminate manner subject to a requirement of strict necessity. Further, the directive also does not preclude the retention of data beyond statutory data retention periods when strictly necessary to shed light on serious criminal offences or attacks on national security, when the offences or attacks have already been established, or if their existence may reasonably be suspected. Real-time data may also be used when it is limited to people in respect of whom there is a valid reason to suppose that they are involved in terrorist activities. Such use of data must be subject to prior review by an independent body to ensure that real-time collection is limited to what is strictly necessary. The Court notes that in urgent cases that the review should take place promptly (presumably rather than after the event).
Finally, a national court may not apply a provision of national law empowering it to limit the temporal effects of a declaration of illegality which declaration the national court must make in respect of national legislation due to incompatibility with the e-Privacy Directive, and evidence obtained illegally should not be relied on in court.
The common theme across the cases was the acceptability of the retention and analysis of communications data generally. The Court has re-iterated its general approach, unsurprisingly linking – as the Advocate General also did – between the Privacy International ruling and that in La Quadrature du Net. In its approach, the Court relied generously on its previous rulings, which demonstrates that there is quite a thick rope of cases, all to broadly the same effect. While the Court based its ruling on the ePrivacy directive (which is specific to communications and communications data), it also based its ruling more generally on Articles 7 and 8 of the Charter. It is noteworthy that the Court did not just refer to its case law on communications data but also to the Canada PNR opinion, underlining that there is a similar approach no matter the type of data in issue. The Court also relied on Schrems II, implicitly confirming aspects of its approach there and embedding that decision in its jurisprudence. The underlying concern in Schrems II was the same as here: that is, data collected by private actors are accessed by state actors. In sum, even in the interests of national security, general and indiscriminate surveillance does not satisfy the test of strict necessity and proportionality. While its general approach might be similar to what has gone before, there are still some points of interest and new ground covered.
The IPT seems to have been the only court amongst those making references that still has not accepted that the retention of data falls within the scope of the e-Privacy Directive, relying on the reasoning of the Court on the Data Protection Directive in relation to passenger name records in an early case. In addition to re-establishing the well-trodden principles regarding the impact of requiring electronic service providers to retain data bringing the entire scheme within scope of the e-Privacy Directive, and different functions of Article 1(3) (scope of directive) and Art 15 (derogation from directive), the Court took the opportunity to say something about the scope of the GDPR, the successor legislation to the Data Protection Directive. In effect, the Court has stopped the line of reasoning found in that early PNR judgment – it cannot be used to determine the scope of the GDPR which should be understood in line with Art 1(3) of the e-Privacy Directive.
The Court has emphasised a couple of aspects of the legal regime surrounding surveillance that are worth a second look. Firstly, while the Court says nothing about the form of law on which a surveillance may be based, in its analysis of Article 52(1) Charter it does say that the same law must contain the constraints. The principle then has wider application than just communications data. This raises questions about forms of surveillance rolled out by the police based on broad common law powers, or – as in the recent Bridges decision – in a mix of legislation, common law and code. These sorts of surveillance – although in public – may also give rise to a feeling of being subject to constant surveillance, though the Court’s jurisprudence on video-surveillance under the GDPR has not yet grappled with this issue. It may be, however, that the Court would take a different view on the extent to which ‘private life’ would be engaged in such circumstances. It is also worth noting that the views of the independent body must be binding on the SIAs; this reiterates the point that in principle approval must be sought in advance.
The Court also made clear that the rights in issue are not just privacy and data protection; it specifically referred here to freedom of expression and flagged the distinctive of those under professional duties of confidentiality (doctors, lawyers) and whistleblowers. It did not, however, consider whether any infringement was justified in this context. The list of possible rights affected is not limited to freedom of expression: in Schrems II the Court highlighted the right to a remedy. It is not inconceivable that the right to association could also be affected. Presumably the same points of analysis apply – that general and indiscriminate monitoring cannot be justified even in the interests of national security. The Court also recognised, in La Quadrature du Net, the positive obligations on the State in relation to Article 3 and 8 ECHR and the corresponding article in the Charter – Articles 4 and 7. The balancing of these positive obligations provided the framework for the Court’s analysis of types of surveillance that did not immediately fall foul of its prohibition of of general and indiscriminate data retention. In this context, it might almost be said that the Court is reformulating public interest objectives (such as national security or the fight against sexual abuse of children) as positive obligations and thus bringing them in a rights balancing framework.
The Court’s reasoning in both cases also gave us some insight into the meaning of national security. It is distinct from and covers more serious issues that the objectives listed in Art 15. While this in principle seems to allow more intrusive measures to be justified, it seems that the Court has limited the circumstances of when it can be used. It does not overlap seemingly with those grounds in Article 15 e Privacy Directive. So, even might be argued reading this part of the judgments that serious crime cannot be blurred with national security. The devil will be in the detail here, a tricky one for any independent body to patrol – and in terms of permitted surveillance it is not clear what the consequences in practice would be.
The headline news, however, must be the ruling of the Court relating to measures that do not fall within the prohibition as general and indiscriminate measures. This on one level is not totally novel; it is implied, for example, in Tele2/Watson, para 106. The questions relate to what level of generality of surveillance would be permissible, and in relation to what sort of objective? Para 137 seems to limit targeted retention of communications data to matters of national security (including terrorism), but the Court then wheels out the same reasoning in relation to serious crime and public safety, and seems to envisage similar safeguards in both cases. This then means that the test of ‘strict necessity’ is doing a lot of work in distinguishing between the legitimate and illegitimate use of surveillance measures. The Court has historically not been particularly strong on what it requires of a necessity test – let alone one requiring strict necessity – in other cases involving the interference with Charter rights.
The final point relates to the procedural questions. The Court was clear that striking down incompatible law cannot have suspended effect. Yet, that is precisely what the English court did in Watson when allowing the UK government several months to get its house in order. The Court of Justice also held here that illegally obtained evidence cannot be used in court, relying on the need to ensure that the rights granted by EU law are effective. While the status of EU law in the British courts may currently be uncertain on the face of it this might mean that convictions based on data between the handing down of Tele2/Watson, or at latest its application by the English courts, until the revision of the regime might be open to challenge whatever the domestic rules on evidence might say. Of course, even if we did not have to deal with the jurisprudential consequences of Brexit, the Court of Appeal, in its approach to Tele2/Watson ignored the aspects of the judgment directed at Tele2 referring court despite the fact that element of the judgment was an interpretation of EU law having general application, so it is to be assumed that still more would it ignore a ruling in a different case altogether.
This post first appeared on the EU Law Analysis blog and is reproduced here with permission and thanks.