Admissibility of Social Media Evidence: A Case Study

In the Internet age, people still have the same interests and passions as they had before electronic communications became pervasive, but they have different methods of expressing them. It may be a challenge to apply traditional rules of law to those methods. This note reviews one example of such a challenge, with respect to the use of evidence from Facebook and the reliance on Wikipedia to inform the tribunal of relevant facts.

In Landry c. Provigo Québec Inc (Maxi & Cie), 2011 QCCLP 1802 (CanLII), Madame Landry complained about harassment at her workplace at Provigo, the Quebec grocery chain. Her case was dismissed by the Commission de la santé et de la sécurité du travail [Workplace Health and Safety Commission] and went on appeal to the Commission des lésions professionnelles [Commission on workplace injuries]. The complainant relied in part on printouts of comments made on the Facebook page by friends of a colleague whom she had recently friended. These comments were not favourable to her. (The actual nature of the harassment is not disclosed in the judgment.)


The employer objected to the admission of the Facebook pages on three grounds: they lacked integrity, they were hearsay, and their admission would violate privacy rights of the people commenting. The Commission noted that the use of Facebook pages in labour cases seemed to be well established, citing several cases, but that the admissibility had apparently not been directly challenged before.(para 33)


The complainant had printed out several pages of her friend’s Facebook account in four batches, but submitted only some of the pages as evidence, the others not being about her, she said. The employer claimed that the selection was the equivalent of untraceable editing and the text was therefore unreliable. The Commission held that the texts submitted themselves had integrity, in that nothing had been omitted from them or changed in them. There was no parallel to edited surveillance videos or other instances where evidence as presented might have been altered.

There are logically two aspects of ‘integrity’ here, though neither the employer nor the Commission distinguished them consistently. The first, and perhaps less important, is the impact of the editing, dealt with above. The second, more fundamental, is the authentication of the evidence. The Civil Code of Quebec (article 2855) says that any material evidence must be shown independently to be authentic – unless it is an electronic document (what Quebec legislation refers to as a ‘technological document’). Technological documents benefit from a special rule (and a printout of a text taken from a website is such a document for this purpose.) In that context, the Commission reviewed Quebec’s Act to Establish a Legal Framework for Information Technology, RSQ, c C-1.1

Section 5 of the Act says, in part:

 The legal value of a document, particularly its capacity to produce legal effects and its admissibility as evidence, is neither increased nor diminished solely because of the medium or technology chosen.

A document whose integrity is ensured has the same legal value whether it is a paper document or a document in any other medium, insofar as, in the case of a technology-based document, it otherwise complies with the legal rules applicable to paper documents.

A document in a medium or based on technology that does not allow its integrity to be confirmed or denied may, depending on the circumstances, be admissible as testimonial evidence or real evidence and serve as commencement of proof, as provided for in article 2865 of the Civil Code.

The Civil Code says that a technological document needs to be authenticated only if it falls into the third paragraph of section 5 of the Act, i.e. it is one ‘based on a technology that does not allow its integrity to be confirmed or denied’. However, section 7 of the Act presumes that such a document has integrity and puts the onus of proof on the party denying it. Section 7 says this:

It is not necessary to prove that the medium of a document or that the processes, systems or technology used to communicate by means of a document ensure its integrity, unless the person contesting the admission of the document establishes, upon a preponderance of evidence, that the integrity of the document has been affected.

Further, the Code of Civil Procedure (s. 89) requires for this purpose that the party challenging the document submit an affidavit specifying the facts and reasons that support the lack of integrity of the document in question. This had not been done in the Landry case. (para 52)

By way of comparison, the Uniform Electronic Evidence Act creates a presumption of the integrity of the computer system of a proponent of an electronic record, for the purposes of the best evidence rule. (Uniform Act s. 5(a)) It may be that this presumption could be used in support of authentication as well. (Uniform Act s. 2(2)) It is certainly a good deal weaker than the presumption in Quebec law.

The Commission noted finally that it was not bound by the strict rules of evidence in any event. (para 45) It held that the challenge on the ground of lack of integrity failed.


The employer said that the words of Mme Landry’s fellow employees were hearsay and thus inadmissible. The employees were not available for cross-examination by the employer. The Commission disposed of this argument quickly. The names and pictures of the employees in question appeared with their comments on the Facebook pages printed and submitted by the complainant. If the employer wanted to call them to the hearing and ask them questions, it was free to do so.

The Commission found as well (para 56) that the documents presented ‘sufficient guarantees of trustworthiness’, though it did not explain how. Perhaps the presumptions of integrity already set out were considered to spill over into the hearsay analysis.

The complainant submitted as well that the comments of the employees were not hearsay because not proferred to prove their truth but to show that they had been made by her fellow workers and constituted harassment. The Commission could have decided the point in her favour on this ground as well.

(The status of the content of websites as hearsay can be a problem in many cases. A more detailed review of the issue will have to await another occasion.)


The employer pointed to the rules of procedure of the Commission itself, which say this:

11. [...] Il doit toutefois, même d’office, rejeter tout élément de preuve obtenu dans des conditions qui portent atteinte aux droits et libertés fondamentaux et dont l’utilisation est susceptible de déconsidérer l’administration de la justice. [It must however, even on its own motion, reject any element of the evidence obtained in conditions that infringe on fundamental rights and liberties and whose use may tend to bring the administration of justice into disrepute.] (para 11)

One might compare Manitoba’s Privacy Act, s. 7:

No evidence obtained by virtue or in consequence of a violation of privacy in respect of which an action may be brought under this Act is admissible in any civil proceedings.

Thus the Commission needed to decide if there were any such infringement. Are comments posted on Facebook part of the private sphere? (para 65). It held that they were not. Getting access to the comments was done in accordance with the general rules of Facebook. Becoming someone’s friend gave access to comments of friends of that friend. While that relationship lasted, the complainant had legitimate access to the comments. (After the complaint was filed, she was defriended and lost access to them.) ‘This principle of interaction of different users is at the base of a social network like Facebook.” (para 67)[translation]

While comments on Facebook are personal to the commenter and do not bind anyone else, one must distinguish between personal comments and private comments. (para 69)

[70] Une personne qui détient un compte Facebook permet à ses amis et aux amis de ses amis de prendre connaissance de ses commentaires. Cette personne peut contrôler la liste de ses amis, mais il devient plus difficile de contrôler l’accès à son profil aux amis de ses amis, liste qui peut s’allonger presque à l’infini. Nous sommes donc loin du caractère privé du profil de cette personne et des commentaires qu’elle émet. [A person who has a Facebook account allows his or her friends and friends of friends to know his or her comments. The person may control the list of his or her friends, but it is more difficult to control access to one's profile by friends of friends, a list that can stretch almost to infinity. We are thus far from any private character with respect to the person's profile and the comments that he or she makes.]

One may compare Leduc v Roman 2009 CanLII 6838 (ON SC) in Ontario:

Facebook profiles are not designed to function as diaries; they enable users to construct personal networks or communities of ‘friends’ with whom they can share information about themselves, and on which ‘friends’ may post information about the user. (para 31)

Thus the Commission was in the Canadian mainstream in treating Facebook comments are not private for the purpose of admissibility into evidence. (Slaw has discussed frequently the separate question whether personal information on Facebook is compellable on examination for discovery. The general answer is Yes, but fishing expeditions for such information will be discouraged.)


The Commission began its review by considering the nature of Facebook, on which it informed itself by referring to the French edition of Wikipedia. It said that it was common practice for tribunals to refer to dictionaries to understand the usual meaning of a word or expression. ‘In the modern world’, Wikipedia could be considered a dictionary for this purpose. (para 31).

Slaw has considered the use of Wikipedia in courts and tribunals here , here and here . Despite considerable scepticism about such use, including in courts themselves, references are remarkably common, especially, it appears, in administrative tribunals. Nicolas Vermeys and Patrick Gingras classify 184 Canadian uses found on CanLII up to the summer of 2011: to define medical, legal and economic terms (and at least once to understand if a term was defamatory), to explain scientific, mechanical or mathematical concepts and the rules of grammar, to get a translation, or to provide background facts of history, geography, economics and sociology. (‘Citer ou ne pas citer: la preuve par Wikipedia’, Repères, September 2011, Editions Yvon Blais., text and notes 8 to 21 and 79 to 83. The article was inspired by the Landry case discussed here.)


The Commission in Landry did not question whether Wikipedia could be considered a dictionary. However, the site defines itself as a kind of encylopedia, that is, an explanatory text rather than a definition. It is in the nature of explanations to be debatable, and thus more risky to rely on than a source of definitions. (We can skip over the rich literature on statutory and other forms of legal interpretation that raise issues with the use of definitions themselves.)

As an explanation, how reliable is Wikipedia? Its texts can be created and modified at any time by anyone with access to the Internet (with a few exceptions now for particularly notorious topics or controversial people whose entries are more tightly controlled.) Thus the level of expertise involved can vary widely, and from one day to the next. Wikipedia itself says this about reliance:


Some information on Wikipedia may create an unreasonable risk for readers who choose to apply or use the information in their own activities or to promote the information for use by third parties.

Perhaps even more important than the question of knowledgeability of the contributors is that of their neutrality. The text of articles can be slanted to one’s own preferences. While the whole world may be watching (and amending), and Wikipedia has editors to enforce its standards of neutrality of point of view and of citation of authoritative sources, any particular article may escape their attention for an indefinite time. (Extra caution would be needed for a printed version of a Wikipedia page, which might be created immediately after a tendentious amendment, before a balance was restored.)

This risk has been faced frequently in immigration cases, where the political climate of a refugee claimant’s homeland can be critical to the determination of refugee status. Tribunals have been sceptical of Wikipedia entries in this context, whether used by claimants or by adjudicators. Jahazi v Canada (Citizenship and Immigration) , 2010 FC 242 (CanLII) para 60, but cf Moiseev v Canada (Citizenship and Immigration) , 2008 FC 88 (CanLII) para 24. In a non-judicial context, Professor Vermeys mentions an incident where a student made extensive changes to Wikipedia’s page on the novel Brave New World in order to cover up his plagiarism of the original material on the page.

Sometimes Wikipedia articles seem to be submitted as a kind of principal evidence, but often they are an appeal to judicial notice. Judicial notice may be taken of facts that are so well known as not to be disputed, or so amenable to clear proof from readily accessible and authoritative sources that direct proof in the courtroom may be dispensed with. While online texts may well have such a character, it is far from clear that Wikipedia is among them. A well-known study comparing Wikipedia’s reliability favourably to that of the Encyclopedia Britannica has been doubted in later reviews.

It is probably appropriate to take Wikipedia’s disclaimer at face value. “DO NOT RELY UPON ANY INFORMATION FOUND IN WIKIPEDIA WITHOUT INDEPENDENT VERIFICATION. ”

Limits and Protections

Given these widely-known concerns, why is Wikipedia used so often? Clearly its accessibility and the very large number of articles work in its favour. In many cases it is arguable that the content used is quoted more for its economy of expression than for any novelty in the meaning. In other words, the court or tribunal knows the commonly accepted meaning and used Wikipedia only as a confirmation and expression of it. If the Wikipedia entry ran contrary to the court’s prior understanding, it would almost certainly not be used. For this purpose, its use is fairly safe.

Again, this kind of use provides background information, it sets the context for the principal dispute. It seems much riskier to rely on Wikipedia for the decisive point in a legal proceeding. Again in its own words, “If you need specific advice (for example, medical, legal, financial, marital or risk management) please seek a professional licensed or knowledgeable in that area.”

Vermeys and Gingras also point out the usefulness of the references to source material set out in Wikipedia articles. Wikipedia’s editors do check that the references are independent of the contributor, and of serious reputation. Thus the site is a good starting point for finding the type of authority that it is safe to rely on, and it may help orient the search for the material too.

It appears in some cases that administrative tribunals have allowed themselves more latitude about the strict rules of evidence, including proof of reliability of their sources, than would courts. The Commission notes this distinction on another point in the Landry decision and has relied on it in other decisions. That seems a hard distinction to generalize, however. People’s rights are decided in both types of body, and the parties should be able to rely on the decisions being made on a credible basis.

At the very least, anyone thinking of giving legal effect to a Wikipedia article should be aware of the ‘metadata’ of the article, namely the Discussion page in which differences of interpretation are sometimes debated by contributors or editors, and the History page that shows when changes have been made to the article and by whom. It is sometimes possible to know the actual author of an amendment, and sometimes only the IP address from which the amendment came. Further research can sometimes track down a likely author, but with more effort and resources than a court or tribunal, or the parties to a proceeding, may wish to devote to the effort. The need to refer to these pages is mentioned by the Montreal Municipal Court in R. v Cianfagna 2007 CanLII 25904 (QC C.M.)(reversed on appeal on other grounds). (Wikipedia’s Risks page cited earlier mentions these steps as well.)


The admissibility of information from Facebook seems common across Canada. The types of questions faced by the Commission in Landry may need to be disposed of in individual cases, as examples of more general principles of evidence like authentication and hearsay. Whether the presumption of reliability of ‘technological documents’ provided by the Quebec legal framework statute is appropriate invokes a wider debate than that of the present article.

As for Wikipedia, the Commission in Landry skated over the big issues, though arguably without doing any harm in that case. The characteristics of Facebook relied on by the Commission are themselves widely known and might qualify for judicial notice without the intermediary of Wikipedia. The challenge to any particular description of Facebook may come from the frequency with which Facebook itself changes its features and terms of use, rather than from any potential amendments to the Wikipedia article on it.

ENDNOTE on Jurisprudential Solitudes

This article relies heavily on reasoning and sources assembled by a handful of Quebec jurists who have produced a remarkably broad and deep literature on electronic communications and the law, principally in that province but with a good understanding of other Canadian sources and international thinking.

Besides the article mentioned above about the Landry decision and Wikipedia, one may note as a mere selection of recent sources:

  1. Vermeys and Gingras, ‘La Preuve issue des médias sociaux : Capture, préparation, présentation ‘ (Legal IT, April 2011)
  2. Gingras and Vermeys, Technologies de l’information En bref (No. 1, May 2012), notably ‘Qu’est-ce que Facebook? Plaidoyer pour une qualification juridique des médias sociaux’
  3. Vermeys and Gingras, ‘Je tweet, tu clavardes, il blogue : les aléas juridiques de la communication électronique‘ (Barreau du Québec 2011). The second half deals with admissibility of evidence from social media.
  4. Gingras and Vincent Gautrais, La preuve des documents technologiques (2010) 22 Les Cahiers de la propriété intellectuelle, num. 2, updated in 2012 for the Barreau du Québec.

Professor Gautrais, Pierre Trudel and several colleagues have just launched a web site devoted to the Act to provide a Legal Framework for Information Technology on its tenth anniversary, setting out interpretation and case law. The site promises to be of great interest. The differences between Quebec law and that of the rest of Canada on these topics would support an additional article or series of articles.

Retweet information »


  1. The Commission sur les lésions professionnelles (CLP) has returned to the social media admissibility question in Campeau et Services alimentaires Delta Dailyfood Canada inc., 2012 QCCLP 7666 (CanLII).

    A worker was injured and had to take a lot of time off work. At one point her injuries caused her a case of depression that also kept her off work. To test whether this was serious, the employer created a fictitious account on Facebook, giving the alleged member characteristics chosen deliberately to appeal to the employee. The employee friended the fictitious member, which gave that member – and the employer – access to the employee’s friends-only posts and information.

    The workplace injuries tribunal rejected the employee’s case, and the employee appealed to the CLP. The CLP held that the evidence from Facebook was not admissible. They reviewed article 2858 of the Quebec Civil Code, which says (as above) that a tribunal must reject evidence obtained in conditions that infringe fundamental rights and liberties and whose use would cast the administration of justice in disrepute. This article is an express exception to the admissibility of any fact relevant to the dispute.

    The decision analyses these factors in some detail (paragraphs 15 through 76). The infringement of rights affected the employee’s privacy rights, guaranteed against private-sector infringement by the Quebec Charter though not the federal Charter.

    The test for bringing the administration of justice into disrepute was a matter of proportionality. Judged by standards from cases in the Quebec Court of Appeal and the Supreme Court of Canada, the CLP held that using the FB evidence wouuld have that result. Fishing expeditions seem likely to meet that standard consistently.

    Would the case be decided differently without the express rule in the Civil Code about infringement of rights and the reputation of the administration of justice? Without a private-sector privacy law extending to employment relationships? Could a tribunal in a common-law jurisdiction have reached the same result?