Advertise on Durhampost.ca
2M Views and counting!

AI, deepfakes and divorce in the age of fake evidence

You get a text message. One spouse sends a screenshot that shows the other spouse admits to concealing funds. Later, a threatening voicemail follows. Or maybe there is a video that seems to depict behaviour that could be relevant to a parenting dispute.

About two years ago, the most likely question would have been: “Was that really sent?”

Nowadays, we have another option: “Is that really real?”

With the development of artificial intelligence, creating images that appear to be photos, mimicking voices, and generating communications that look genuine has never been easier. With deepfakes, which were previously hard to produce without a great deal of technical expertise, it is now possible to use these technologies using applications and/or software purchased through online retailers.

Therefore, for divorcing couples, lawyers, and judges, this presents a difficult issue. While more family law disputes rely upon digital evidence than ever before, it is becoming increasingly simple to alter or destroy digital evidence.

When Screenshots Are No Longer the Only Reason for Disputes

Screenshots are fast becoming a standard feature of modern litigation. There are so many ways in which text messages, WhatsApp discussions, email exchanges, social media posts and others can prove the specifics of a family law dispute, including financial disclosure and the manner in which divorced parents communicate with each other.

Electronic evidence has long been subject to regulation within Canadian courts.

Ontario’s Evidence Act places the obligation upon the party presenting an electronic document as evidence to provide sufficient proof that the document is what he/she/they purport it to be. This requirement for verification of authenticity appears in similar form under the Canada Evidence Act.

This does not mean that all screenshots require forensic examination.

In the 2025 Ontario Court of Appeal decision regarding a screenshot of text messages, the court noted that while the threshold for verifying the authenticity of a document is generally low, testimony from the individual who produced the screenshot was sufficient to establish its authenticity. However, questions about potential tampering with the screenshot could still influence the weight assigned to the evidence.

AI will significantly increase the number of disputes over screenshots.

The Criminal Law Perspective

When a person’s liberty is at risk based on what appears on a screen, developing technology to convincingly manufacture digital evidence poses serious concerns for criminal defence lawyer Manbir Nirwal and founder of Nirwal Law in Brampton.

Digital communications such as screenshots, surveillance footage, photographs, social media messages, recordings of conversations and voice messages frequently serve as important pieces of evidence in criminal trials. A single digital communication may support the majority of the Crown’s case.

“The concern here is that people naturally have confidence in things they can see/hear. Once you play back a recording that sounds just like someone else speaking, your immediate assumption tends to be that it is actually them. AI forces us to examine our assumptions and scrutinize far more closely where the evidence originated.”

The importance of knowing where the digital document originated from may thus increase. Lawyers may wish to know who produced the document, where it was saved/stored, how it was acquired by them, and what occurred to it prior to arriving at court.

Additionally, context is crucial.

A screenshot depicting three messages from an otherwise lengthy conversation that contains hundreds of messages may convey a vastly different interpretation than the entire conversation. AI adds yet another level of complexity since lawyers may eventually begin questioning whether the conversation even existed.

Nirwal states that digital evidence should always be evaluated as part of a larger context rather than being viewed exclusively as irrefutable evidence.

“A screenshot can present an overwhelmingly compelling appearance. However, appearances constitute only one element of your evaluation. You will need to evaluate the totality of surrounding evidence and assess both the origin/source of materials presented and if there exists credible reason to doubt their reliability.”

What If Someone Said “That’s Not My Voice”?

Another possible problem for Gurasish Pal Sandhu is caused by voice cloning technology.

Some AI systems currently require only several seconds of spoken dialogue to clone an individual’s voice. Therefore, voice-cloned communications could be manufactured to appear as though a person had uttered words they did not intend to utter.

Brampton-based criminal defence lawyer Gurasish Pal Sandhu of GPS Criminal Law Firm believes this increases the significance of additional evidence in evaluating the reliability of digital material when parties dispute material via digital means.

“If a party maintains that a recording has been fabricated, then merely assessing the validity of the recording will not suffice. It may be necessary to review the device from which the recording purportedly originated, examine the original file(s), review circumstances related to creation of the recording and other evidence supporting or opposing it.”

The conflict may occur in numerous contexts.

Consider a scenario in which police receive a voice message in which someone apparently threatens them. The recipient recognizes the voice. The accused denies having recorded it.

Questions of credibility are no longer solely dependent on credibility. Technology itself becomes part of the dispute.

Ontario courts have already addressed assertions that electronic communications can be fabricated. In recent decisions, courts have stated that a mere suspicion that fabrication has occurred does not render electronically generated evidence inadmissible. Rather, determining authenticity, reliability and worth of consideration remain separate questions.

The sophistication of these arguments will likely grow as AI develops further.

Divorce in the Age of Deepfakes

Divorce in family court may be particularly affected by this trend, since spouses typically bring extensive digital records of their relationship into the courtroom.

An aggressive text could serve as a basis for a parental rights dispute. Communications may also be used to contest or verify a spouse’s version of an agreement. Social media posts could refute sworn testimony given in court. Emails may represent important documentation if spouses are unable to agree on financial arrangements during a separation.

Manjeet Kaur, a Brampton family lawyer and founder of Kaur Family Law suggests that individuals undergoing a separation should be cautious against treating every screenshot as definitive evidence.

“Separation cases can be emotionally charged and sometimes parties feel compelled by one specific message proving everything they believe. Context matters greatly in terms of digital evidence and with increasing ease in manipulating it through AI, preserving original documents and surrounding communications may become even more critical.”

Additionally, there are practical considerations.

Divorced partners often have access to large volumes of digital content about each other, e.g., hundreds of pictures taken by each partner, videos shot at family gatherings, and years’ worth of voice messages. These are essentially ideal materials for producing synthetic media that looks authentic.

“As technology advances and provides greater opportunities for manufacturing fake digital media, formerly married spouses will face increased concern about availability/accessibility of vast quantities of digital data involving each other. Ultimately, regardless of how much access spouses have to each other’s digital lives, courts will continue to decide what constitutes valid/credible evidence. Therefore, clients should take extra care to preserve their original files/documentation when disputes occur.”

So What Should You Save?

The simple act of taking a screenshot will likely continue to exist but it’s unlikely that simply capturing a screenshot (or other compressed versions) will remain as useful as the full source document/material when at all possible.

Taking the entire text-based exchange via texting/messaging app provides better context than a single screenshot. In many cases, the original video/photograph/audio files may include information missing from the compressed version. Emails may contain metadata. When necessary to prove something, devices themselves may also need to be considered if there are disputes about authenticity.

We should consider how we modify/crop/re-save/share screenshots of digital documents prior to sharing them with others for legal purposes, etc.

None of this indicates Canadian citizens can automatically dismiss damaging evidence as being a “deep fake.”

Rather than “seeing” being the best way to know what happened, it will now be a part of proving that is occurred.

You May Also Like To Read

First works start on Bowmanville Hospital expansion

Ajax says ‘remove your stuff’ for street lighting replacements

Any more victims of builder who duped homeowner of $56K?

Ontario Shores gets $2mn for specialized trauma, PTSI care

Help create a fund for students facing financial barriers

Leave a Reply