Concept
Authenticating Electronic Evidence
Authentication and identification
Authenticating Electronic Evidence
Electronic evidence is authenticated under the same rule as every other exhibit. Rule 901(a) speaks of “an item of evidence” and draws no line by medium, so the proponent of a web page, a social-media profile, a text message, or an e-mail must produce evidence sufficient to support a finding that the item is what the proponent claims it is.1 The medium changes which facts are available to prove the connection; it does not change the standard. The proponent’s claim sets what must be connected, the measure is whether a reasonable juror could find the connection, and the judge screens the foundation while the jury decides genuineness. A foundation that would suffice for a paper letter therefore suffices for an e-mail of the same character, and a foundation that would fail for a letter does not improve because the item arrived on a screen.
The claim a proponent makes about an electronic item is usually a claim about a person: that she created the page, wrote the post, or sent the message. The link between the item and that person is attribution, and attribution is where these cases are fought. It is hard because electronic material is cheap to produce and easy to alter, and because anyone can open an account or build a profile in another person’s name. A page can display a real person’s name, photograph, and life details without that person ever having seen it. Taken alone, those details are like a letterhead or a signature block: an impostor copies them first, so they reveal little about who made the item. The proponent has to point to something beyond the item’s face.
Those risks lead some to ask for a stricter foundation for online material than for paper. The prevailing approach keeps the ordinary standard and treats the danger of forgery as part of two familiar questions: whether the proponent’s evidence lets a reasonable juror find attribution, and, once it does, how much the jury should believe it. Forged paper has always been possible, and Rule 901(a) answers that possibility by requiring connecting evidence and leaving the contest over genuineness to the jury once such evidence exists. A higher threshold for electronic items would move that contested question from the jury to the judge.
Two settings recur, and they pose different problems. Public content, such as a page anyone can visit or a post on a social-networking site, raises the question of who stood behind the account, because a viewer can describe what appeared on the screen but usually cannot say who put it there. Private messages sent to a recipient start from a better position, since the recipient can testify about receiving them and about the account and device involved, and the dispute centers on what the device, the account, and the content together show about the sender.
Web Pages and Social Media
A printout of a web page must meet the same test as any other exhibit: evidence sufficient to support a finding that it is what the proponent claims it is.2 What sets web pages apart is the claim usually made about them. A profile page or a post matters in a lawsuit only if some person stands behind it, so the proponent ordinarily claims that a particular person created the page or answers for what it says, and the foundation must reach that claim. The Second Circuit’s decision in Zhyltsou, which Maryland’s highest court cites as Vayner, shows why.3 The government first told the trial court that it offered a Russian social-networking page only as what a special agent was observing online that day, and the agent offered nothing that tied the page to the defendant.4 Yet the government wanted the page for a different purpose. It helped the prosecution only if the defendant, or someone acting for him, had created it, because only then did it tend to show that he used the moniker “azmadeuz” on Skype and probably also for the Gmail address from which a forged birth certificate was sent, as a cooperating witness claimed.5 Authorship was therefore the claim the proof had to support.6
The gap lay in what the government could show. Everything on the page that pointed to the defendant was information the cooperating witness and likely others also knew, and some of those people may have had reasons to build a page falsely attributed to him.7 Beyond the page itself, nothing suggested that the defendant had a profile on the site at all, and nothing suggested that the site required identity verification to open an account, which would have made attribution less speculative.8 Those absences point to what a proponent can try to supply: a connection that exists outside the page and that a copyist could not easily manufacture. The court declined to say what would have been enough, because the type and quantum of evidence needed to authenticate a web page depends on context.9 But the page was offered to corroborate the cooperating witness’s account, so Rule 901 required some basis beyond that witness’s own testimony on which a reasonable juror could find that it was the defendant’s profile and not just any page on the Internet, and the lack of one required exclusion.10
In Kieffer the government offered archived screenshots of a website it attributed to the defendant; a detective had retrieved them from an Internet archive shortly before trial, and the defendant objected that the exhibit was not authenticated.11 The proponent’s claim had two links: that the screenshots accurately showed the site as it existed during the relevant period, and that the site was the defendant’s. A witness addressed ownership by testifying that a domain name stays under its registrant’s exclusive control until the registration expires, that the defendant registered this domain in May 2004, and that the registration did not expire until May 2010.12 The screenshots identified the defendant as the organization’s executive director, and two other witnesses testified that the exhibit resembled the site they had viewed in 2007, which spoke to the accuracy of the archived copy.13 The court upheld the ruling that the objection went to weight and not admissibility, because the judge need only assess whether the proponent has offered a foundation from which the jury could reasonably find the evidence authentic.14 A page naming the defendant as director would carry little alone, since a page can say anything about its owner. The domain-registration testimony tied the content to him by showing who controlled that address during the period. In Zhyltsou the only proof was the page’s own face.
Platform records and the defendant’s own accounts can supply the outside link as well. In Hassan, which the Maryland court cites as applying the same standard to social-media postings, the government offered screenshots of the defendants’ Facebook pages with certifications from the records custodians of Facebook and Google that the pages and videos were kept as business records. The trial court found Rule 902(11) satisfied, and it separately found Rule 901(a) satisfied by tracing the pages and accounts, through internet protocol addresses, to the defendants’ mailing and e-mail addresses; the Fourth Circuit found no abuse of discretion.15 The two rulings answer different questions. A custodian’s certification speaks to how the company stores its pages, while the address tracing speaks to who was behind the accounts, so a certification alone leaves authorship unproved. That fits the Second Circuit’s observation that none of the Rule 902 categories, which include certain public records, periodicals, and business records, applied to the profile page in Zhyltsou, so the government had no shortcut around extrinsic proof.16 Rule 902 and its certification procedure receive their own sections later in this chapter.
In United States v. Lewisbey the ties piled up around a Facebook page. The defendant admitted the posts were his, which the Seventh Circuit treated as enough, but the court added that the page listed his nickname, birth date, and former city of residence, that the e-mail addresses associated with the account matched the one linked to his iPhone and his former university address, and that more than 100 photos of him appeared, many matching photos on the phone.17 The Facebook application on the phone was linked to the account, and messages on the account described trips to gun shows on dates when shows actually took place at those locations.18 A nickname, birth date, and residence resemble the name, photograph, and biographical details that failed in Zhyltsou. What differs is that these details sat beside facts about the account’s linked e-mail addresses and the defendant’s own phone, which a person copying a public profile would be unlikely to match, and beside content that outside events corroborated.
With both poles in view, the disagreement among courts is easier to state. Griffin v. State, a decision of Maryland’s highest court, involved a printout of a MySpace profile offered to show that the defendant’s girlfriend had written a post reading “snitches get stitches.”19 The profile carried a photograph of her with the defendant, her birth date, her location, and a reference to freeing someone called Boozy, which the dissent identified as a nickname for the defendant.20 The majority held that admitting the printout was an abuse of discretion: given the prospect that someone other than the claimed creator built the profile or posted the comment, a social-networking printout calls for more authentication than a birth date and a photograph.21 Two judges dissented. They read the rule to require only that a reasonable juror could find authenticity, a standard they said federal courts have applied almost unanimously, and they concluded that the photograph, birth date, location, and nickname reference met it, leaving the risk of an impostor to the jury as weight.22
The Second Circuit in Zhyltsou sided with the dissent on the standard. It was skeptical that greater scrutiny is required for evidence from the Internet, but it did not have to decide the question, because the government’s showing failed under the general rule.23 Maryland’s highest court later took the same position in Sublet v. State, a decision framed as an elucidation of Griffin: a trial judge authenticating social-networking evidence asks whether there is proof from which a reasonable juror could find that the evidence is what the proponent claims, and the court embraced the standard as stated in Zhyltsou.24 The facts of Sublet show the standard at work. The defense offered four pages of a Facebook conversation during its cross-examination of a State witness, Parker, and each entry displayed a profile name and picture.25 Parker acknowledged that one profile name was hers but denied writing the comments on the fourth page, and the court held that the trial court did not err in excluding the pages.26
Both sides start from the same worry: a false profile is easy to build. The Griffin majority reasoned that a photograph and a birth date are items anyone can copy, so a printout showing only those says little about who opened the account. A court applying the ordinary standard answers that the reasonable-juror test already screens out a profile that rests on nothing but copyable details, as Zhyltsou shows, and that the chance an impostor made the page is a subject for cross-examination and rebuttal.27 The Griffin majority adds that the trial court must address authentication whether or not anyone has raised a motive to fabricate or manipulate the item.28 The disagreement is not merely verbal. On the Griffin profile, the majority excluded and the dissent would have admitted; on a bare profile like the one in Zhyltsou, the two approaches reach the same result. The usual exam error is treating the defendant’s name and photograph on a page as proof that the defendant created it. A stronger answer states the reasonable-juror standard, notes that some courts, as in Griffin, ask for more, and then identifies the facts outside the page that link it to the claimed author. Private messages, which go to a recipient and are not published for all to see, start from a stronger position and come next.29
Texts and E-mails
Private electronic messages are authenticated under the same standard as every other item: the proponent must produce evidence sufficient to support a finding that the message is what the proponent claims it is.31 The claim is almost always about a person, namely that a particular individual wrote and sent the text or e-mail, so the contest is over origin. Origin is a standard authentication question for any writing, whether the doubt concerns a signature or the account an e-mail came from.32 Texts and e-mails start from a stronger footing than public posts. They travel directly from one party to an intended recipient and are not published for everyone to see, so the recipient, the account, and the device are all identifiable places to look for the sender, while a public page offers little beyond what it says about itself.33 The proof is assembled from circumstances, and courts draw them from four recurring sources: the account and address details on the message, the device and who possessed it, the specifics and nicknames in the contents, and later events that match what the message said.
United States v. Siddiqui shows all four sources working at once. The e-mails bore the defendant’s university address, which matched the address on a message sent to him that his own counsel had introduced, and when a recipient replied, the e-mail system automatically addressed the response to that address.34 The text showed knowledge of the details of his conduct concerning an award and the investigation that followed, apologized for cutting short a visit to a Swiss research institute in a way that matched deposition testimony, and referred to the author as “Mo,” a nickname both recipients recognized as his.35 Both recipients also testified that he telephoned soon after the e-mails arrived and made the same requests, and the Eleventh Circuit found no abuse of discretion in the ruling that the e-mails were authenticated.36
Each of those facts does a different job. An address on a message is a label, and a label is the easiest thing for a sender to claim, so it earns weight only because the same address appeared elsewhere in the defendant’s own correspondence and because the reply function sent responses there. Private details and a nickname known to acquaintances narrow the possible authors to people close to the events. The follow-up call closes the remaining gap: the person who telephoned made the requests the e-mail had made, which is conduct by the claimed sender that matches the message. Other courts have relied on the same kind of matching conduct. In Commonwealth v. Amaral the court reasoned that the defendant’s own actions authenticated e-mails, because one said he would be at a certain place at a certain time and he appeared, and another gave a telephone number that he answered immediately when an investigator called it.37 In Dickens v. State the court found text messages authenticated because the number on one showed it came from the defendant’s phone and others referred to his right to see the couple’s minor child and to their wedding vows.38
When the message sits on a seized phone, the phone itself becomes evidence of who stood behind it. In Lewisbey the government needed only evidence sufficient to support a finding that the text messages were actually sent and received by the defendant.39 The iPhone was taken from him at his arrest, and in a recorded jail call he told his mother that the police had taken his phone, while the Samsung was recovered from a bedroom that both parents identified as belonging exclusively to him.40 The iPhone’s properties described it as “Big Dave’s,” its contacts listed his mother under “Mom” along with the name of his former attorney, both phones listed the Texas Home Depot stores where he used to work, and a confidential informant arranged gun sales with him on the Samsung.41 The Seventh Circuit called that more than enough to establish that the two phones were his, grounding the point in the rule’s allowance for contents and distinctive characteristics taken together with all the circumstances.42 Each fact narrows the field of possible users, and the informant’s use of the Samsung to arrange the sales shows that the phone was the channel through which the deals were actually made.
Siddiqui placed the identifying facts inside the messages, and Lewisbey placed them around the phone. Either arrangement works because the question under Rule 901(a) is whether a reasonable juror could find the connection, and a cluster of independent facts supports that finding where any one of them, such as a sender line, could have been supplied by anyone.43
What if the only link between a message and a person is that it appears on a phone the person admits owning? Commonwealth v. Koch brings that question into focus because the court was divided over it. The Commonwealth offered drug-sale text messages from a phone that the appellee, Koch, acknowledged was hers, and no first-hand testimony from an author or a recipient accompanied them.44 The Commonwealth argued that it did not need to prove authorship at all, because she was charged as both a principal and an accomplice, so the crucial fact was that her physical phone was used in drug transactions.45 Koch answered that the claim being made was that she sent and received the messages on her personal phone, that mere possession of a phone does not prove authorship of messages sent from it, and that corroboration such as content unique to the parties was needed. She added that the detective had conceded someone else likely wrote at least some of the messages.46
The Pennsylvania court observed that it had not yet addressed how text messages may be authenticated without first-hand testimony from an author or recipient, stressed that the authentication burden is low, called the question close, and treated authorship as a potentially relevant part of the analysis.47 It then described a split among other high courts. Nevada’s court in Rodriguez v. State, which involved messages sent from a victim’s phone that had been stolen from her during an attack, held that when a text message is challenged the proponent must explain the purpose for which it is offered and provide direct or circumstantial corroborating evidence of authorship, and other high courts likewise call for corroboration.48 Arizona’s court in State v. Forde took a position closer to that of the trial court in Koch, finding the burden met where both sender and recipient were registered subscribers of phone numbers and both possessed the phones used to send and receive.49 Because the court in Koch was divided, the decision is best read as a map of the dispute and not as a settled rule, and it would be a mistake to report it as holding either that possession of a phone suffices or that it never does.
The split exists because possession answers a narrower question than authorship. A phone is a portable object that people lend, share, leave on tables, and lose, so having it shows who could have typed a message and leaves open who did. The Nevada facts show the gap, since the messages there were sent from a phone stolen from its owner during an attack, a setting where ownership of the device said nothing about who typed. Corroboration closes the gap by tying the words to a particular person, through information that specifies the sender, events before or after the message that match it, or other facts signifying that the message is what the proponent claims, which are the circumstances the Pennsylvania court listed.50 The registered-subscriber reasoning in Forde adds a second fact to bare possession, a link between the number and a person, which is why it reaches further than a rule resting on possession alone. The proponent’s claim sets the target here as it does everywhere under Rule 901(a). The Commonwealth’s argument that authorship need not be shown was an attempt to narrow the claim to use of the phone, while the appellee characterized the claim as messages she herself sent and received.
The pair shows how an exam answer should proceed. The first step is to find the facts beyond possession and the sender line, and the second, if there are none, is to state both positions and the reason the stricter one worries about shared devices. The usual error is to treat a sender address or the phone’s owner as proof of authorship without asking what else ties the words to the person.
A hearsay objection raises a different question from an authentication objection, and the two have to be kept apart. In Siddiqui the defendant’s lawyer objected that the government had not reliably shown who sent the e-mail, and the Eleventh Circuit held that objections of that kind go to authentication and are not hearsay objections.51 The court added that the e-mails the defendant sent were his own admissions under Rule 801(d)(2)(A), while exchanges with a recipient on unrelated matters came in as non-hearsay to show the pair’s relationship and custom of communicating by e-mail.52 In Lewisbey the texts the defendant sent were his own statements, and the messages he received were admitted for context and not for their truth.53 The two analyses connect at one point, because the finding that the defendant wrote a message is what makes it the defendant’s own statement. Authentication is still only the first gate, and a message written by someone else and offered for its truth needs its own route past the hearsay rule.