EXECUTIVE SUMMARY OF A STUDY FOR THE EP LIBE COMMITEE.
FULL TEXT ACCESSIBLE HERE
by Mirja GUTHEIL, Quentin LIGER, Aurélie HEETMAN, James EAGER, Max CRAWFORD (Optimity Advisors)
Hacking by law enforcement is a relatively new phenomenon within the framework of the longstanding public policy problem of balancing security and privacy. On the one hand, law enforcement agencies assert that the use of hacking techniques brings security, stating that it represents a part of the solution to the law enforcement challenge of encryption and ‘Going Dark’ without systematically weakening encryption through the introduction of ‘backdoors’ or similar techniques. On the other hand, civil society actors argue that hacking is extremely invasive and significantly restricts the fundamental right to privacy. Furthermore, the use of hacking practices pits security against cybersecurity, as the exploitation of cybersecurity vulnerabilities to provide law enforcement with access to certain data can have significant implications for the security of the internet.
Against this backdrop, the present study provides the LIBE Committee with relevant, actionable insight into the legal frameworks and practices for hacking by law enforcement. Firstly, the study examines the international and EU-level debates on the topic of hacking by law enforcement (Chapter 2), before analysing the possible legal bases for EU intervention in the field (Chapter 3). These chapters set the scene for the primary focus of the study: the comparative analysis of legal frameworks and practices for hacking by law enforcement across six selected Member States (France, Germany, Italy, the Netherlands, Poland and the UK), with further illustrative examples from three non-EU countries (Australia, Israel and the US) (Chapter 4). Based on these analyses, the study concludes (Chapter 5) and presents concrete recommendations and policy proposals for EU action in the field (Chapter 6).
The international and EU-level debates on the use of hacking techniques by law enforcement primarily evolve from the law enforcement challenge posed by encryption – i.e. the ‘Going Dark’ issue.
‘Going Dark’ is a term used “to describe [the] decreasing ability [of law enforcement agencies] to lawfully access and examine evidence at rest on devices and evidence in motion across communications networks”.1
According to the International Association of Chiefs of Police (IACP), law enforcement agencies are not able to investigate illegal activity and prosecute criminals without this evidence. Encryption technologies are cited as one of the major barriers to this access. Although recent political statements from several countries (including France, Germany, the UK and the US) seemingly call for ‘backdoors’ to encryption technologies, support for strong encryption at international and EU fora remains strong. As such, law enforcement agencies across the world started to use hacking techniques to bypass encryption. Although the term ‘hacking’ is not used by law enforcement agencies, these practices essentially mirror the techniques used by hackers (i.e. exploiting any possible vulnerabilities – including technical, system and/or human vulnerabilities – within an information technology (IT) system).
Law enforcement representatives, such as the IACP and Europol, report that access to encrypted and other data through such hacking techniques brings significant investigative benefits. However, it is not the only possible law enforcement solution to the ‘Going Dark’ issue. Outside of the scope of this study, the other options include: requiring users to provide their password or decrypt their data; requiring technology vendors and service providers to bypass the security of their own products and services; and the systematic weakening of encryption through the mandated introduction of ‘backdoors’ and/or weakened standards for encryption.
With the benefits of hacking established, a 2016 Joint Statement published by the European Union Agency for Network and Information Security (ENISA) and Europol2 noted that the use of hacking techniques also brings several key risks.
The primary risk relates to the fundamental right to privacy and freedom of expression and information, as enshrined in international, EU and national-level law. Hacking techniques are extremely invasive, particularly when compared with traditionally intrusive investigative tools (e.g. wiretapping, house searches etc.). Through hacking, law enforcement can gain access to all data stored or in transit from a device; this represents a significant amount of data (e.g. a recent investigation by Dutch law enforcement collected seven terabytes of data, which translates into around 86 million pages of Microsoft Word documents3), as well as extremely sensitive data (e.g. a person’s location and movements, all communications, all stored data etc.). Consequently, the use of hacking techniques will inherently restrict the fundamental right to privacy.
Therefore, current debates at international and EU fora focus on assessing and providing recommendations on the current legal balances and safeguards for the restriction of the right to privacy by hacking techniques. However, these debates have assumed that hacking practices are necessary for law enforcement and simply require governing laws; they have not discussed whether the use of hacking techniques by law enforcement is necessary and proportional. The law enforcement assertions regarding the necessity of these invasive tools have not been challenged.
The second key risk relates to the security of the internet. Law enforcement use of hacking techniques has the potential to significantly weaken the security of the internet by “[increasing] the attack surface for malicious abuse”4. Given that critical infrastructure and defence organisations, as well as law enforcement agencies themselves, use the technologies targeted and potentially weakened by law enforcement hacking, the potential ramifications reach far beyond the intended target.
As such, debates at international and EU fora focus on the appropriate balances between security and privacy, as well as security and cybersecurity. Regarding security v. privacy, the debates to date have assessed and provided recommendations on the legislative safeguards required to ensure that hacking techniques are only permitted in situations where a restriction of the fundamental right to privacy is valid in line with EU legislation (i.e. legal, necessary and proportional). Regarding security v. cybersecurity, the debates have been limited and primarily centre around the use and/or reporting of zero-day vulnerabilities discovered by law enforcement agencies.
Further risks not discussed in the Joint Statement but covered by this study include: the risks to territorial sovereignty – as law enforcement agencies may not know the physical location of the target data; and the risks related to the supply and use of commercially-developed hacking tools by governments with poor consideration for human rights.
Alongside the analysis of international and EU debates, the study presents hypotheses on the legal bases for EU intervention in the field. Although possibilities for EU legal intervention in several areas are discussed, including mutual admissibility of evidence (Art. 82(2) TFEU), common investigative techniques (Art. 87(2)(c) TFEU), operational cooperation (Art. 87(3) TFEU) and data protection (Art. 16 TFEU, Art. 7 & 8 EU Charter), the onus regarding the development of legislation in the field is with the Member States. As such, the management of the risks associated with law enforcement activities is governed at the Member State level.
As suggested by the focus of the international and EU discussions, concrete measures need to be stipulated at national-level to manage these risks. This study presents a comparative analysis of the legal frameworks for hacking by law enforcement across six Member States, as well as certain practical aspects of hacking by law enforcement, thereby providing an overview of the primary Member State mechanisms for the management of these risks. Further illustrative examples are provided from research conducted in three non-EU countries.
More specifically, the study examines the legal and practical balances and safeguards implemented at national-level to ensure: i) the legality, necessity and proportionality of restrictions to the fundamental right to privacy; and ii) the security of the internet.
Regarding restrictions to the right to privacy, the study first examines the existence of specific legal frameworks for hacking by law enforcement, before exploring the ex-ante and ex-post conditions and mechanisms stipulated to govern restrictions of the right to privacy and ensure they are legal, necessary and proportional.
It is found that hacking practices are seemingly necessary across all Member States examined, as four Member States (France, Germany, Poland and the UK) have adopted specific legislative provisions and the remaining two are in the legislative process. For all Member States except Germany, the adoption of specific legislative provisions occurred in 2016 (France, Poland and the UK) or will occur later (Italy, the Netherlands). This confirms the new nature of these investigative techniques.
Additionally, law enforcement agencies in all Member States examined have used, or still use, hacking techniques in the absence of specific legislative provisions, under so-called ‘grey area’ legal provisions. Given the invasiveness of hacking techniques, these ‘grey area’ provisions are considered insufficient to adequately protect the right to privacy.
Where specific legal provisions have been adopted, all stakeholders agree that a restriction of the right to privacy requires the implementation of certain safeguards. The current or proposed legal frameworks of all six Member States comprise a suite of ex-ante conditions and ex-post mechanisms that aim to ensure the use of hacking techniques is proportionate and necessary. As recommended by various UN bodies, the provisions of primary importance include judicial authorisation of hacking practices, safeguards related to the nature, scope and duration of possible measures (e.g. limitations to crimes of a certain gravity and the duration of the hack, etc.) and independent oversight.
Although many of these types of recommended conditions are common across the Member States examined – demonstrated in the below table – their implementation parameters differ. For instance, both German and Polish law permit law enforcement hacking practices without judicial authorisation in exigent circumstance if judicial authorisation is achieved in a specified timeframe. However, the timeframe differs (three days in Germany compared with five days in Poland). These differences make significant difference, as the Polish timeframe was criticised by the Council of Europe’s Venice Commission for being too long.5
Furthermore, the Member States examined all accompany these common types of ex-ante and ex-post conditions with different, less common conditions. This is particularly true for ex-post oversight mechanisms. For instance, in Poland, the Minister for internal affairs provides macro-level information to the lower (Sejm) and upper (Senat) chambers of Parliament;6 and in the UK, oversight is provided by the Investigatory Powers Commissioner, who reviews all cases of hacking by law enforcement, and the Investigatory Powers Tribunal, which considers disputes or complaints surrounding law enforcement hacking.7
|Key ex-ante considerations|
|Judicial authorisation||The legal provisions of all six Member States require ex-ante judicial authorisation for law enforcement hacking. The information to be provided in these requests differ.
Select Member States (e.g. Germany, Poland, the UK) also provide for hacking without prior judicial authorisation in exigent circumstances if judicial authorisation is subsequently provided. The timeframes for ex-post authorisation differ.
|Limitation by crime and duration||All six Member States restrict the use of hacking tools based on the gravity of crimes. In some Member States, the legislation presents a specific list of crimes for which hacking is permitted; in others, the limit is set for crimes that have a maximum custodial sentence of greater than a certain number of years. The lists and numbers of years required differ by Member State.
Many Member States also restrict the duration for which hacking may be used. This restriction ranges from maximum 1 month (France, Netherlands) to a maximum of 6 months (UK), although extensions are permitted under the same conditions in all Member States.
|Key ex-post considerations|
|Notification and effective remedy||Most Member States provide for the notification of targets of hacking practices and remedy in cases of unlawful hacking.|
|Reporting and oversight||Primarily, Member States report at a micro-level through logging hacking activities and reporting them in case files.
However, some Member States (e.g. Germany, Poland and the UK) have macro-level review and oversight mechanisms.
Furthermore, as regards the issue of territoriality (i.e. the difficulty law enforcement agencies face obtaining the location of the data to be collected using hacking techniques), only one Member States, the Netherlands, legally permits the hacking of devices if the location is unknown. If the device turns out to be in another jurisdiction, Dutch law enforcement must apply for Mutual Legal Assistance.
As such, when aggregated, these provisions strongly mirror Article 8 of the European Convention on Human Rights, as well as the UN recommendations and paragraph 95 of the ECtHR judgement in Weber and Saravia v. Germany. However, there are many, and varied, criticisms when the Member State conditions are examined in isolation. Some of the provisions criticised include: the limits based on the gravity of crimes (e.g. the Netherlands, France and Poland); the provisions for notification and effective remedy (e.g. Italy and the Netherlands); the process for screening and deleting non-relevant data (Germany); the definition of devices that can be targeted (e.g. the Netherlands); the duration permitted for hacking (e.g. Poland); and a lack of knowledge amongst the judiciary (e.g. France, Germany, Italy and the Netherlands).With this said, certain elements, taken in isolation, can be called good practices. Such examples are presented below.
Select good practice: Member State legislative frameworks
Germany: Although they were deemed unconstitutional in a 2016 ruling, the provisions for the screening and deletion of data related to the core area of private life are a positive step. If the provisions are amended, as stipulated in the ruling, to ensure screening by an independent body, they would provide strong protection for the targeted individual’s private data.
Italy: The 2017 draft Italian law includes a range of provisions related to the development and monitoring of the continued use of hacking tools. As such, one academic stakeholder remarked that the drafting of the law must have been driven by technicians. However, these provisions bring significant benefits to the legislative provisions in terms of supervision and oversight of the use of hacking tools. Furthermore, the Italian draft law takes great care to separate the functionalities of the hacking tools, thus protecting against the overuse or abuse of a hacking tool’s extensive capabilities.
Netherlands: The Dutch Computer Crime III Bill stipulates the need to conduct a formal proportionality assessment for each hacking request, with the assistance of a dedicated Central Review Commission (Centrale Toetsings Commissie). Also, the law requires rules to be laid down on the authorisation and expertise of the investigation officers that can perform hacking.
With these findings in mind, the study concludes that the specific national-level legal provisions examined provide for the use of hacking techniques in a wide array of circumstances. The varied combinations of requirements, including those related to the gravity of crimes, the duration and purpose of operations and the oversight, result in a situation where the law does not provide for much stricter conditions than are necessary for less intrusive investigative activities such as interception.
Based on the study findings, relevant and actionable policy proposals and recommendations have been developed under the two key elements: i) the fundamental right to privacy; and ii) the security of the internet.
Recommendations and policy proposals: Fundamental right to privacy
It is recommended that the use of ‘grey area’ legal provisions is not sufficient to protect the fundamental right to privacy. This is primarily because existing legal provisions do not provide for the more invasive nature of hacking techniques and do not provide for the legislative precision and clarity as required under the Charter and the ECHR.
Furthermore, many of these provisions have only recently been enacted. As such, there is a need for robust evidence-based monitoring and evaluation of the practical application of these provisions. It is therefore recommended that the application of these new legal provisions is evaluated regularly at national level, and that the results of these evaluations are assessed at EU-level.
If specific legislative provisions are deemed necessary, the study recommends a range of good practice, specific ex-ante and ex-post provisions governing the use of hacking practices by law enforcement agencies. These are detailed in Chapter 6.
Policy proposal 1: The European Parliament should pass a resolution calling on Member States to conduct a Privacy Impact Assessment when new laws are proposed to permit and govern the use of hacking techniques by law enforcement agencies. This Privacy Impact Assessment should focus on the necessity and proportionality of the use of hacking tools and should require input from national data protection authorities.
Policy proposal 2: The European Parliament should reaffirm the need for Member States to adopt a clear and precise legal basis if law enforcement agencies are to use hacking techniques.
Policy proposal 3: The European Parliament should commission more research or encourage the European Commission or other bodies to conduct more research on the topic. In response to the Snowden revelations, the European Parliament called on the EU Agency for Fundamental Rights (FRA) to thoroughly research fundamental rights protection in the context of surveillance. A similar brief related to the legal frameworks governing the use of hacking techniques by law enforcement across all EU Member States would act as an invaluable piece of research.
Policy proposal 4: The European Parliament should encourage Member States to undertake evaluation and monitoring activities on the practical application of the new legislative provisions that permit hacking by law enforcement agencies.
Policy proposal 5: The European Parliament should call on the EU Agency for Fundamental Rights (FRA) to develop a practitioner handbook related to the governing of hacking by law enforcement. This handbook should be intended for lawyers, judges, prosecutors, law enforcement officers and others working with national authorities, as well as nongovernmental organisations and other bodies confronted with legal questions in the areas set out by the handbook. These areas should cover the invasive nature of hacking techniques and relevant safeguards as per international and EU law and case law, as well as appropriate mechanisms for supervision and oversight.
Policy proposal 6: The European Parliament should call on EU bodies, such as the FRA, CEPOL and Eurojust, to provide training for national-level members of the judiciary and data protection authorities, in collaboration with the abovementioned handbook, on the technical means for hacking in use across the Member States, their potential for invasiveness and the principles of necessity and proportionality in relation to these technical means.
Recommendations and policy proposals: Security of the internet
The primary recommendation related to the security of the internet is that the position of the EU against the implementation of ‘backdoors’ and similar techniques, and in support of strong encryption standards, should be reaffirmed, given the prominent role encryption plays in our society and its importance to the EU’s Digital Agenda. To support this position, the EU should ensure continued engagement with global experts in computer science as well as civil society privacy and digital rights groups.
The actual impacts of hacking by law enforcement on the security of the internet are yet unknown. More work should be done at the Member State level to assess the potential impacts such that these data can feed in to overarching discussions on the necessity and proportionality of law enforcement hacking. Furthermore, more work should be done, beyond understanding the risks to the security of the internet, to educate those involved in the authorisation and use of hacking techniques by law enforcement.
At present, the steps taken to safeguard the security of the internet against the potential risks of hacking are not widespread. As such, the specific legislative provisions governing the use of hacking techniques by law enforcement, if deemed necessary, should safeguard the security of the internet and the security of the device, including reporting the vulnerabilities used to gain access to a device to the appropriate technology vendor or service provider; and ensure the full removal of the software or hardware from the targeted device.
Policy proposal 7: The European Parliament should pass a resolution calling on Member States to conduct an Impact Assessment to examine the impact of new or existing laws governing the use of hacking techniques by law enforcement on the security of the internet.
Policy proposal 8: The European Parliament, through enhanced cooperation with Europol
and the European Union Agency for Network and Information Security (ENISA), should
reaffirm its commitment to strong encryption considering discussions on the topic of hacking by law enforcement. In addition, the Parliament should reaffirm its opposition to the implementation of ‘backdoors’ and similar techniques in information technology infrastructures or services.
Policy proposal 9: Given the lack of discussion around handling zero-day vulnerabilities, the European Parliament should support the efforts made under the cybersecurity contractual Public-Private Partnership (PPP) to develop appropriate responses to handling zero-day vulnerabilities, taking into consideration the risks related to fundamental rights and the security of the internet.
Policy proposal 10: Extending policy proposal 4, above, the proposed FRA handbook should also cover the risks posed to the security of the internet by using hacking techniques.
Policy proposal 11: Extending policy proposal 5, training provided to the judiciary by EU bodies such as FRA, CEPOL and Eurojust should also educate these individuals on the risks posed to the security of the internet by hacking techniques.
Policy proposal 12: Given the lack of discussion around the risks posed to the security of the internet by hacking practices, the European Parliament should encourage debates at the appropriate fora specific to understanding this risk and the approaches to managing this risk. It is encouraged that law enforcement representatives should be present within such discussions.