The Council of Europe Cybercrime Committee's (T-CY) recent decision to approve new international rules for law enforcement access to user data without strong privacy protections is a blow for global human rights in the digital age. The final version of the draft Second Additional Protocol to the Council of Europe’s (CoE) widely adopted Budapest Cybercrime Convention, approved by the T-CY drafting committee on May 28th, places few limits on law enforcement data collection. As such, the Protocol can endanger technology users, journalists, activists, and vulnerable populations in countries with flimsy privacy protections and weaken everyone's right to privacy and free expression across the globe.
The Protocol now heads to members of CoE's Parliamentary Committee (PACE) for their opinion. PACE’s Committee on Legal Affairs and Human Rights can recommend further amendments, and decide which ones will be adopted by the Standing Committee or the Plenary. Then, the Council of Ministers will vote on whether to integrate PACE's recommendations into the final text. The CoE’s plan is to finalize the Protocol's adoption by November. If adopted, the Protocol will be open for signatures to any country that has signed the Budapest Convention sometime before 2022.
The next step for countries is at the signature stage. Countries can present a signature without reservation as to ratification, acceptance, or approval. At that time, they can reserve the right not to abide by certain provisions in the Protocol, in particular Article 7 on direct cooperation between law enforcement and companies holding user data.
If countries sign the Protocol as it stands and in its entirety, it will reshape how state police access digital data from Internet companies based in other countries by prioritizing law enforcement demands, sidestepping judicial oversight, and lowering the bar for privacy safeguards.
CoE’s Historical Commitment to Transparency Conspicuously Absent
While transparency and a strong commitment to engaging with external stakeholders have been hallmarks of CoE treaty development, the new Protocol’s drafting process lacked robust engagement with civil society. The T-CY adopted internal rules that have fostered a largely opaque process, led by public safety and law enforcement officials. T-CY’s periodic consultations with external stakeholders and the public have lacked important details, offered short response timelines, and failed to meaningfully address criticisms.
In 2018, nearly 100 public interest groups called on the CoE to allow for expert civil society input on the Protocol’s development. In 2019, the European Data Protection Board (EDPB) similarly called on T-CY to ensure “early and more proactive involvement of data protection authorities” in the drafting process, a call it felt the need to reiterate earlier this year. And when presenting the Protocol’s draft text for final public comment, T-CY provided only 2.5 weeks, a timeframe that the EDPB noted “does not allow for a timely and in-depth analysis” from stakeholders. That version of the Protocol also failed to include the explanatory text for the data protection safeguards, which was only published later, in the final version of May 28, without public consultation. Even other branches of the CoE, such as its data protection committee, have found it difficult to provide meaningful input under these conditions.
Last week, over 40 civil society organizations called on CoE to provide an additional opportunity to comment on the final text of the Protocol. The Protocol aims to set a new global standard across countries with widely varying commitments to privacy and human rights. Meaningful input from external stakeholders including digital rights organizations and privacy regulators is essential. Unfortunately, CoE's answer kindly refused, and will likely finally adopt the Treaty in November 2021.
With limited incorporation of civil society input, it is perhaps no surprise that the final Protocol places law enforcement concerns first while human rights protections and privacy safeguards remain largely an afterthought. Instead of attempting to elevate global privacy protections, the Protocol’s central safeguards are left largely optional in an attempt to accommodate countries that lack adequate protections. As a result, the Protocol encourages global standards to harmonize at the lowest common denominator, weakening everyone’s right to privacy and free expression.
Eroding Global Protection for Online Anonymity
The new Protocol provides few safeguards for online anonymity, posing a threat to the safety of activists, dissidents, journalists, and the free expression rights of everyday people who go online to comment on and criticize politicians and governments. When Internet companies turn subscriber information over to law enforcement, the real-world consequences can be dire. Anonymity also plays an important role in facilitating opinion and expression online and is necessary for activists and protestors around the world. Yet the new Protocol fails to acknowledge the important privacy interests it places in jeopardy and, by ensuring most of its safeguards are optional, permits police access to sensitive personal data without systematic judicial supervision.
As a starting point, the new Protocol’s explanatory text claims that: "subscriber information … does not allow precise conclusions concerning the private lives and daily habits of individuals concerned,” deeming it less intrusive than other categories of data.
This characterization is directly at odds with growing recognition that police frequently use subscriber data access to identify deeply private anonymous communications and activity. Indeed, the Court of Justice of the European Union (CJEU) recently held that letting states associate subscriber data with anonymous digital activity can constitute a ‘serious’ interference with privacy. The Protocol’s attempt to paint identification capabilities as non-intrusive even conflicts with CoE’s own European Court of Human Rights (ECtHR). By encoding the opposite conclusion in an international protocol, the new explanatory text can deter future courts from properly recognizing the importance of online anonymity. As the ECtHR held doing so would, “deny the necessary protection to information which might reveal a good deal about the online activity of an individual, including sensitive details of his or her interests, beliefs and intimate lifestyle.”
Articles 7 and 8 of the Protocol in particular adopt intrusive police powers while requiring few safeguards. Under Article 7, states must clear all legal obstacles to “direct cooperation” between local companies and law enforcement. Any privacy laws that prevent Internet companies from voluntarily identifying customers to foreign police without a court order are incompatible with Article 7 and must be amended. “Direct cooperation” is intended to be the primary means of accessing subscriber data, but Article 8 provides a supplementary power to force disclosure from companies that refuse to cooperate. While Article 8 does not require judicial supervision of police, countries with strong privacy protections may continue relying on their own courts when forcing a local service provider to identify customers. Both Articles 7 and 8 also allow countries to screen and refuse any subscriber data demands that might threaten a state’s essential interests. But these screening mechanisms also remain optional, and refusals are to be “strictly limited,” with the need to protect private data invoked only in “exceptional cases.”
By leaving most privacy and human rights protections to each state’s discretion, Articles 7 and 8 permit access to sensitive identification data under conditions that the ECtHR described as “offer[ing] virtually no protection from arbitrary interference … and no safeguards against abuse by State officials.”
The Protocol’s drafters have resisted calls from civil society and privacy regulators to require some form of judicial supervision in Articles 7 and 8. Some police agencies object to reliance on the courts, arguing that judicial supervision leads to slower results. But systemic involvement of the courts is a critical safeguard when access to sensitive personal data is at stake.
The Office of the Privacy Commissioner of Canada put it cogently: “Independent judicial oversight may take time, but it’s indispensable in the specific context of law enforcement investigations.”
Incorporating judicial supervision as a minimum threshold for cross-border access is also feasible. Indeed, a majority of states in T-CY’s own survey require prior judicial authorization for at least some forms of subscriber data in their respective national laws.
At a minimum, the new Protocol text is flawed for its failure to recognize the deeply private nature of anonymous online activity and the serious threat posed to human rights when State officials are allowed open-ended access to identification data. Granting states this access makes the world less free and seriously threatens free expression. Article 7’s emphasis on non-judicial ‘cooperation’ between police and Internet companies poses a particularly insidious risk, and must not form part of the final adopted Convention.
Imposing Optional Privacy Standards
Article 14, which was recently publicized for the first time, is intended to provide detailed safeguards for personal information. Many of these protections are important, imposing limits on the treatment of sensitive data, the retention of personal data, and the use of personal data in automated decision-making, particularly in countries without data protection laws. The detailed protections are complex, and civil society groups continue to unpack their full legal impact. That being said, some shortcomings are immediately evident.
Some of Article 14’s protections actively undermine privacy—for example, paragraph 14.2.a prohibits signatories from imposing any additional “generic data protection conditions” when limiting the use of personal data. Paragraph 14.1.d also strictly limits when a country’s data protection laws can prevent law enforcement-driven personal data transfers to another country.
More generally, and in stark contrast to the Protocol’s lawful access obligations, the detailed privacy safeguards encoded in Article 14 are not mandatory and can be ignored if countries have other arrangements in place (Article 14.1). States can rely on a wide variety of agreements to bypass the Article 14 protections. The OECD is currently negotiating an agreement that might systematically displace the Article 14 protections and, under the United States Clarifying Lawful Overseas Use of Data (CLOUD) Act, the U.S. executive branch can enter into “agreements” with other states to facilitate law enforcement transfers. Paragraph 14.1.c even contemplates informal agreements that are neither binding, nor even public, meaning that countries can secretly and systematically bypass the Article 14 safeguards. No real obligations are put in place to ensure these alternative arrangements provide an adequate or even sufficient level of privacy protection. States can therefore rely on the Protocol’s law enforcement powers while using side agreements to bypass its privacy protections, a particularly troubling development given the low data protection standards of many anticipated signatories.
The Article 14 protections are also problematic because they appear to fall short of the minimum data protection that the CJEU has required. The full list of protections in Article 14, for example, resembles that inserted by the European Commission into its ‘Privacy Shield’ agreement. Internet companies relied upon the Privacy Shield to facilitate economic transfers of personal data from the European Union (EU) to the United States until the CJEU invalidated the agreement in 2020, finding its privacy protections and remedies insufficient. Similarly, clause 14.6 limits the use of personal data in purely automated decision-making systems that will have significant adverse effects on relevant individual interests. But the CJEU has also found that an international agreement for transferring air passenger data to Canada for public safety objectives was inconsistent with EU data protection guarantees despite the inclusion of a similar provision.
These and other substantive problems with the Protocol are concerning. Cross-border data access is rapidly becoming common in even routine criminal investigations, as every aspect of our lives continues its steady migration to the digital world. Instead of baking robust human rights and privacy protections into cross-border investigations, the Protocol discourages court oversight, renders most of its safeguards optional, and generally weakens privacy and freedom of expression.