In December 2019, Danielle Citron and Bobby Chesney published “Deep Fakes: A Looming Challenge for Privacy, Democracy, and National Security” in the California Law Review. At the time of the paper’s publication, it provided a gaze into the future – a sobering prognosis regarding the social and political consequences artificial intelligence tools being able to generate fake audio, photo and video. The paper (which remains the 147th most cited paper ever within the Social Science Research Network’s library) was uniquely prescient, and is a recommended read for all civil litigators, regardless of their practice area.
Coined in the paper is the concept of the “The Liar’s Dividend” – the perverse benefit accruing to dishonest individuals through the increasing availability of synthetic media. The dividend is twofold: it is not just the problem of fraudulent “evidence” being more easily available to bad actors. It is also that as lawyers, courts, and the public become more aware that audio, video and photos can be convincingly fabricated with technology, the less likely they are to trust authentic evidence – accepting the denials of a bad actor who denies previously damning evidence as “fake”. In estate litigation, this problem is particularly pernicious, because the post-mortem absence of the testator elevates the importance of secondary evidence, such as photos or audio recordings, the reliability of which are now threatened.
Photo and video evidence have never been treated as being infallible – even video intended to record a testator’s intent can be argued as evidence of the opposite, given its limited scope as a snapshot in time. An example of this can be found in Carinci v. Carinci, 2023 ONSC 6094, in which a video recording of a Will’s remote execution (ironically provided by the Respondent) was a key piece in allowing the Applicant’s will challenge to surmount the minimum evidentiary threshold. Hull & Hull’s blog has previously commented on the role and admissibility of audio evidence, as was reflected in the decision of Rudin-Brown et al v. Brown, 2021 ONSC 3366.
Given the critical role of audio and video recordings in these cases, as well as many others, it bears asking: what is to be done when all such evidence is subject to necessary skepticism in the age of AI?
Several categories of evidence commonly relied upon in estate litigation remain largely unaffected by the evidence risk of AI:
Solicitor’s Files: Where available, the contemporaneous records of the drafting solicitor remain critically important evidence. A well-documented solicitor’s file will contain notes of the instructions given to the solicitor, copies of written or e-mail correspondence received from the testator, as well as the solicitor’s own observations with respect to the testator’s ability to instruct and execute a Will, as the Ontario Court of Appeal discussed in Hall v. Bennett Estate, 2003 CanLII 7157. Information in the solicitor’s file has the dual advantage of (usually) predating the testamentary conflict, and having been assessed through the impartial lens of a drafting solicitor, making it an important source of independent information.
Witness Recollections: Though an imperfect source of information, the recollections of family and friends with respect to a testator’s intentions will be an especially important source of information in litigation.
Institutional Records: Documents such as bank statements, e-mail records, and medical records are resistant to the second-order effects of synthetic evidence because they are maintained by independent, third-party institutions, and cannot be created independently by any one person. While the information in these records is less directly connected to a testator’s estate planning, they can still provide a constellation of facts that can be inferentially determinative.
In addition to these sources of evidence, Courts will look to logical consistency within the estate plan. In cases concerning interpretation, for example, the modern approach to the armchair rule permits surrounding circumstances to be considered from the outset of the Court’s analysis, inviting broader considerations of testamentary consistency.
With respect to assessing individual pieces of evidence within estate litigation, a similar assessment of that evidence should be undertaken – does the purported media ‘fit’ with the balance of the evidence, or is it significantly incongruent, despite having the appearance of authenticity? Given the ease with which synthetic media can now be created, anomalous evidence that is challenged as being fraudulent must be put to a great amount of scrutiny.
Thanks for reading!