The Privacy Commissioner of Canada has issued his findings in relation to the investigation of multiple complaints by Canadians regarding the collection, use and disclosure of their personal information by a company based in Romania. The company, Globe24h operates a website which it describes as a “global database of public records”. This global database contains a substantial number of decisions from Canadian courts and administrative tribunals. Some of this content was acquired by scraping court or tribunal websites, or websites such as CanLII. (I wrote about this situation earlier here.)
The problem, from a privacy point of view is that many court and tribunal decisions contain a great deal of personal information. For example, a decision from a divorce case might provide considerable detail about personal assets. Immigration or refugee determination hearings similarly might reveal sensitive personal information. As Commissioner Therrien noted in his findings, the “highly detailed, highly sensitive personal information” found in the decisions that were the focus of the complaints in this case “could have negative reputation impacts (including financial information, health information, and information about children)” (at para 27). Globe24h offers a fee-based service for removal of personal information. A number of the complainants in this case had paid up to 200 euros to have their information removed from decisions in the database.
The Romanian company responded to the investigation by arguing that the Office of the Privacy Commissioner of Canada had no jurisdiction over its activities; and that if it did, Canada’s Personal Information Protection and Electronic Documents Act did not apply because it was engaged in journalistic activities. Alternatively, they argued that they were making use of publicly available information, for which consent is not required under PIPEDA. In this admittedly long blog post, I look at a number of different issues raised in the Commissioner’s findings. You can jump ahead if you like to: Open courts principle and privacy; Extended territorial jurisdiction; Journalism exception; Publicly available personal information; or Crown copyright – the unspoken issue.
The open courts principle – which provides transparency for the justice system in Canada – dictates that decision-makers provide reasons for their decisions and that these decisions be publicly accessible. In the old days, decisions were published in law reports or made available for consultation at court offices. Either way, anyone interested in a particular case had to make some effort to track it down. Decisions were indexed according to subject matter, but were not easily searchable by individual names. The capacity to make court decisions publicly available on the Internet has dramatically increased the ability of the public to access court decisions (and, given the high cost of legal services and the growing number of self-represented litigants, it is not a moment too soon). However, public availability of court decisions on the Internet can raise significant privacy issues for individuals involved in litigation. There is a big difference between accepting that a court decision in one’s case will be published in the interests of transparency and having one’s personal information sucked up and spit out by search engines as part of search results unrelated to the administration of justice.
The main response to this problem to date (from the Canadian Judicial Council’s 2005 Model Policy for Access to Court Records) has been for courts to require the use of technological measures on court websites (and on websites such as CanLII) to prevent search engines from indexing the full text of court decisions. This means that those searching online using a particular individual’s name would not find personal details from court proceedings caught up in the search results. However, these licence terms are only imposed on entities such as CanLII. The general copyright licences on court websites place no such restrictions on the reproduction and use of court decisions. Of course, placing restrictions on the searchability/usability of published decisions can also be a barrier to their innovative reuse. A better approach – or at least a complementary one – might be to be more restrained in the sharing of personal information in published decisions. This latter approach is one recommended by the Office of the Privacy Commissioner of Canada for administrative tribunals. It is unevenly adopted by courts and tribunals in Canada.
While the open courts principle and how Canadian courts and tribunals implement it are relevant to the problem in this case, the Commissioner’s decision does not address these issues. The complaints focussed on the activities of the Romanian company and not on how courts and tribunals manage personal information. Nevertheless, this issue is, to a large extent, at the heart of the problem in this case.
Under basic international law principles, countries cannot apply their laws outside of their own borders. So how could Canadian law apply to a Romanian company’s activities? The answer lies in what some co-authors of mine and I call extended territorial jurisdiction. This arises where activities outside a country’s borders are nonetheless closely connected to that country. After receiving over 20 complaints from Canadians regarding the hosting of their personal information on the Globe24h website, the Privacy Commissioner chose to apply Canada’s PIPEDA to the Romanian company. He did so on the basis that the company was collecting, using and disclosing personal information in the course of commercial activities (key triggers for PIPEDA’s application) and that its activities had a “real and substantial connection” to Canada. This connection was found in the fact that the company chose to include Canadian court and tribunal decisions in its database; that it sourced this material from websites located in Canada; that it accepted requests from Canadians to remove their personal information from its databases; and that it charged Canadians a fee to perform this service. While the company would be subject to Romanian data protection law in general, the Commissioner did not see this as an impediment to applying Canadian law in the specific circumstances. He noted that “It is commonplace in today’s global environment that organizations with an online presence may be subject to data protection laws in multiple jurisdictions depending on the nature of their activities.” (at para 100)
This approach is consistent with that taken by the Office of the Privacy Commissioner of Canada since the Federal Court handed down its decision on this point in Lawson v. Accusearch Inc. Of course, taking jurisdiction over a party in another country and being able to enforce outcomes in accordance with Canadian law are separate matters. In any event, the Privacy Commissioner is relatively toothless even within Canada; in the case of offshore companies any positive results depend largely upon a respondent’s willingness to cooperate with investigations and to change their practices with some gentle nudging. In this case, there seems to be a change of practice on the part of Globe24h, although the extent and durability of this change remain to be seen.
I have previously written about the rather broad and open-ended exception to the application of PIPEDA to the collection, use or disclosure of personal information for “journalistic purposes”. Journalism is capable of a fairly broad interpretation, and in an era of disintermediated information and commentary, a broad approach to this exception is warranted. This may be even more so the case given the Supreme Court of Canada’s recent admonition that privacy laws must be balanced with the freedom of expression. However, an overly broad approach could exclude large swaths of activity from the scope of PIPEDA.
In this case, Globe24h argued that by providing a database of legal information it was entitled to benefit from the journalistic purposes exception. The Commissioner adopted a definition of “journalism” put forward by the Canadian Association of Journalism (CAJ). According to this definition journalism is an activity that has as its goal the communication of information, in a format that has “an element of original production” and that “provides clear evidence of a self-conscious discipline calculated to provide an accurate and fair description of facts, opinion and debate at play within a situation.” (at para 52). The definition is interesting, but it may be under inclusive when it comes to balancing freedom of expression and privacy. This remains an open question. Using this definition, the Commissioner found that the database of public records compiled by Globe24h was not journalism. In particular, he was of the view that the purpose of the database was to generate revenue from different means, including charging individuals who wish to have their personal information removed. He also found that the database did not embody the “original production” required in the CAJ’s definition, and concluded that “Globe24h is republishing information already available online through Canadian court and tribunal websites in a manner that enables the information to be located by search engines, which would not otherwise be possible, so as to profit from individuals’ desire to have this practice stop.” (at para 66).
While there may be an argument that this website does not serve journalistic purposes, the analysis here relies heavily upon the Commissioner’s conclusion that the site’s primary motivation is to derive revenue from individuals who are concerned about their privacy. It is not clear whether, without that element, he would have found that the journalism exception applied. The importance of this poorly worded exception – and the potential of narrow interpretations to conflict with the freedom of expression – leaves one wishing for clearer guidance.
Globe24h also argued that it made use of publicly available personal information. PIPEDA expressly permits the collection, use and disclosure of such information without consent so long as it is used for the purposes for which it was collected and made publicly available. According to the Commissioner, the purpose for which the court decisions were made publicly available was “to promote transparency in the judicial system” (at para 93). He also went on to state that “the purpose for publishing court findings online does not include the association of such findings with individuals’ names in online search results.” (at para 92). The point here, I think, is that the search engine indexing shifts uses of this information away from transparency and towards data mining or snooping; the latter are not consistent with the purposes for which the information was made publicly available.
However, it should be noted that in this case, the assessment of purpose drifts into how the information might be accessed or manipulated by third parties –not by the respondent. This is rather tricky territory. It is a kind of secondary liability in the data protection context: court decisions are made publicly available to anyone around the world; the respondent creates a database that aggregates court decisions from multiple jurisdictions and makes them available. In doing so it enhances the searchability of the decisions by freeing them from technological restrictions. Has it done anything to take it outside the exception? Is the possibility that this new searchability might lead to improper uses of the information by others enough to find that the use does not fall within the exception? My point here is that the problem of excessive personal information in published court decisions seems to be pushed onto those who publish this information (and who thus facilitate the open courts principle), rather than resting with the courts who perhaps should be more careful in deciding what personal information is required to serve the open courts principle and what information is not.
In Canada, court and tribunal decisions are covered by Crown copyright. This lies behind the courts’ ability to dictate licence terms to those who publish these decisions. Recent amendments to the Copyright Act also make it an infringement to circumvent technological protection measures on copyright protected works. Had the Romanian website been publishing court decisions in contravention of the user licence provided by court websites or circumventing court-mandated technological protection measures that blocked the indexing of the court decisions by search engines, then the courts themselves might have sought takedown of these materials or insisted upon compliance with their licence terms. These terms, however, do not appear in the licence for federal court decisions, for decisions of Ontario superior courts, or for decisions of the BC Supreme Court – and this is just a sample. Whether courts should use copyright restrictions to protect privacy values is an interesting question, particularly in an era of increasingly open government. Whether it is realistic or feasible to do so is another good question – if it is not then the privacy issues must be addressed at source. In any event, it may be time for the CJC to revisit its digitally archaic 2005 policy.
The individuals affected by Globe24h turned to the Privacy Commissioner for help when they experienced privacy invasions as a result of the company’s activities. They found a sympathetic ear, and the Commissioner may have achieved some results for them. One can ask, though, where the courts and tribunals have been in all of this. As noted earlier, they should take the lead in addressing privacy issues in their decisions. In addition, while Crown copyright may be an anachronism with the potential to limit free speech, as long as the government clings to it in the face of calls for reform it might consider using it on occasion in circumstances such as these, where inadequate measures designed to protect privacy have failed Canadians and something more is required.