In today’s column, I examine a relatively new phenomenon: attorneys using, at trial, a subpoena-obtained ChatGPT or AI chat transcript as evidence. This is an emerging trend that is undoubtedly going to become increasingly popular. That being said, the use of such transcripts requires a radical rethinking of the evidentiary landscape and the role of juror psychology.
As an example of how the use of AI chat transcripts is not necessarily a slam dunk when it comes to bolstering one side of a case, a recent high-profile criminal case heavily leaned into ChatGPT transcripts that suggested the defendant might have had tendencies alluding to the committing of the crime at issue. A plain reading of the transcripts certainly on the surface would seem to have been damaging to the defense’s side of the case.
For various other reasons, the judge ended up declaring a mistrial, and some of the jurors subsequently spoke to the press about their opinion of the case. One juror made it loudly clear that she was upset that the ChatGPT transcripts were introduced into the trial. She indicated that she uses ChatGPT, which has been quite helpful to her, and that taking apart the AI chat of the defendant seemed an inappropriate intrusion and was unfairly being used against him. The big picture legal perspective is that lawyers need to consider carefully how juries might react not only to the content of AI chat transcripts, but also to the very use and exposure of a person’s “private” AI chats.
Let’s talk about it.
This analysis of AI breakthroughs is part of my ongoing Forbes column coverage on the latest in AI, including identifying and explaining various impactful AI complexities (see the link here).
AI And The Law
As a quick background, I’ve been extensively covering and analyzing a myriad of facets regarding the intersection of AI and the law for many years. You can find my writings not only in my Forbes column but also as posted in Bloomberg Law, ABA Law Journal, The National Jurist, The Global Legal Post, Lawyer Monthly, The Legal Technologist, MIT Computational Law Journal, and so on.
There are two major perspectives on the mixture of AI and law:
- (1) Law & AI. The application of laws to the governance and regulation of AI.
- (2) AI & Law. The application of AI to perform legal reasoning.
Thus, you can apply the law to AI, and conversely, you can apply AI to the law. For my big picture overview of both of these exciting and rapidly evolving realms, see my discussion at the link here and the link here.
When it comes to applying the law to AI, the aim is to establish suitable regulations and provide appropriate governance on how AI should be devised and implemented. There are longstanding concerns that AI makers aren’t giving due attention to the ethical ramifications of their wares. Ethical issues are construed as “soft laws” and aren’t as formidable as legally enacted laws, known as “hard laws”. To level the playing field and keep AI makers on the up-and-up, some believe that we need more AI laws.
On the other side of the coin is the application of AI to the law. This consists of using AI to aid legal activities. Lawyers tap into the latest AI to devise legal strategies, brainstorm to find creative legal arguments, draft court filings, and prepare for cases by having the AI pretend to be an able adversary. For my extensive coverage on AI for legal reasoning (AILR), see the link here.
The Use Of AI Chats As Evidence
I had predicted when ChatGPT was first released that we would eventually see AI chats of generative AI and large language models (LLMs) ultimately being used as evidence at trials, see my early-on prediction at the link here. It has been a slower process than I had envisioned, but we are now witnessing a near explosion in leveraging AI chats for both parties to a legal case.
It certainly makes a great deal of sense to try to obtain such records. Most people share their innermost secrets with AI and consider it their private buddy. People routinely fail to read the online licensing agreements associated with the major LLMs. The licensing for ChatGPT, GPT-5, Claude, Grok, Gemini, CoPilot, and other popular AIs are quick to declare that you are not guaranteed strict privacy. Often, the licensing stipulates that the AI maker can inspect any chats, potentially reuse the chats for further training of their AI, and that they will release the chats if compelled by a court order.
Meanwhile, hundreds of millions of users are daily pouring out their hearts to AI. This becomes an immensely rich source of evidentiary fodder. The odds are that something untoward can be found within the AI conversations. Sometimes the indications are a valid sign and provide credence to aid one side of the case. Other times, the chats are potentially taken out of context or purposely misconstrued to try to score points during a legal case.
The Means Versus The Ends
One thorny rub, which notably is not being given much thought by attorneys eager to tap into AI chats as evidence, is that jurors might not favor this new pathway. An attorney is going to assume that jurors will find the AI chats useful as a kind of exposé of what the party was thinking and intending to undertake. A person’s otherwise cloaked or unspoken inner thoughts are made abundantly clear via the AI chat transcripts. Plus, it is served visibly on a silver platter. Certainly, jurors would eat this up.
Not necessarily so.
We must dig deeply into the psychology of jurors to see why.
Juror Personal Feelings About AI
The chances are that one or more of the jurors on a given case are using or have used ChatGPT or another major LLM. These jurors have indubitably formulated their own opinion about generative AI. This opinion will color the nature of the AI chats used as evidence at a trial. Note that jurors are not going to neutrally perceive the chats, and instead, they will bring their personal bias toward the chats themselves.
I’m sure that most attorneys dryly view AI chats as merely an additional form of an ordinary digital record. No big deal. But such attorneys fail to put themselves into the mindset of the jurors. That can be a losing strategy. Whereas the attorney thinks they have the coup de grace, they are actually heading toward a potential train wreck.
A juror might perceive AI chats as equivalent to a private diary or an intimate conversation with a trusted friend. Some jurors might consider AI chats as being the same as conversing with a psychiatrist and believe that this is tightly protected on a confidentiality basis.
Though savvy attorneys know this is a total misconception, the gist is that this is what a juror potentially believes, regardless of whether it is a proper understanding on their part.
Jurors Can React Negatively
When AI chats are presented with a flourish at trial, some jurors are going to react with shock and dismay. How did those get collected? Why is it legal to do this? They are flabbergasted and will mentally go out of sorts.
In the back of their mind, they are starting to sweat profusely. Well, if it is permissible to openly grab and use AI chats like this, what about the chats I’ve had with AI? Am I at risk? The juror will drift from the trial aspects and become consumed with their own concerns about chatting with AI.
In addition, they are almost surely going to feel that the person in the trial is being ripped off by having their AI chats exposed this way. It is a dastardly, underhanded form of legal trickery. The juror extends sympathy to the person. The juror gets angry at the attorney who has forced this out into the open.
Whatever else the attorney has done in the case will now be highly questioned. If the attorney could do this type of tomfoolery, what other devilish tricks do they have up their sleeve? Worse still, the person whose chats are being torn asunder will get an empathetic reaction from the juror. The content of the AI chats will be viewed and interpreted in the light of a devious means of trying to undercut the person.
Actual Case In Point
This type of juror reaction was made highly visible in a recent court case.
The case was a federal trial that accused a man of starting a horrific blaze that wiped out much of the Los Angeles area known as the Pacific Palisades. You might have heard about the fire since it made headlines in national and international news. This fire ranks as one of the worst wildfires in California history. The destruction was overwhelming, including sadly killing a dozen people, burning down nearly 7,000 structures, and spreading across almost 40 square miles.
At trial, the prosecution made use of AI chats that the defendant had undertaken with ChatGPT. I will say more about this in a moment. The trial eventually landed in the hands of the jury to deliberate. The jury deadlocked. The judge sought to have them reach a verdict but eventually declared a mistrial.
I want to clarify that the AI chats were only a small part of the trial. Do not overweigh them in this instance. Nonetheless, the AI chats were a bone of contention.
Astute Attorneys And AI Chats As Evidence
Attorneys need to formulate a new legal strategy when it comes to making use of compelled AI chats as evidence. Do not assume that AI chats are akin to other everyday digital evidence. Make sure to place yourself in the shoes of the jurors. There might be jurors who will welcome the AI chats, and other jurors who find it revolting that AI chats are being visibly scrutinized for all to see.
Here are my five key recommendations as a leading-edge legal strategy update:
- (1) Be careful about AI chats as evidence. Do not rush pell-mell into using AI chats at trial as though they are the silver bullet that you might think they are; instead, consider the legal value versus the juror repercussions.
- (2) Educate the jurors about AI chats as proper evidence. It might be beneficial to explicitly educate the jurors about how AI chats can be lawfully obtained, and that this is not some improper hacking or surveillance.
- (3) Emphasize the holistic use of AI chats as evidence. Do not make it seem that the whole case turns on the AI chats, which is a dicey ploy, and instead liken the chats to other digital evidence, such as emails, smartphone data, social media postings, etc.
- (4) Detect juror biases about AI chats upfront. During voir dire, attorneys should gently begin probing jurors’ attitudes toward AI chats, such as AI privacy concerns, similar to asking about attitudes toward police, corporations, or social media.
- (5) Be willing to forego using AI chats in court. Lawyers should give due weight to avoiding the introduction of AI transcripts unless they contain uniquely probative admissions that cannot be established through less controversial evidence.
By vigorously employing those recommendations, the evidentiary use of AI chats moves from a wishful knee-jerk choice to a more systemic and legally prudent strategy. That’s what clients expect attorneys to do.
The New Frontier
AI chat transcripts are a new frontier. Unlike previous forms of digital evidence, they imbue a deeply held relationship that many users experience as extraordinarily personal. If jurors increasingly share that perception, attorneys will have to weigh not only whether the evidence is admissible, but also whether introducing it will undermine their credibility with the very people they are trying to persuade. A prudent litigator must treat AI conversations not merely as discoverable evidence, but as evidence that carries an outsized risk of juror backlash.
I will make a bold prediction. The question of AI chats as evidence will be a struggle for a few years. Eventually, it will become the new norm. Jurors will not only be comfortable with AI chats being used in a trial, but they will also come to expect to see them. If AI chats aren’t presented, they will wonder why that is. Did the opposing side believe the AI chats undermined their claims, and therefore are slyly not getting them on the table? It will be quite a change from the era of struggling to employ AI chats as evidence.
A final thought for now. Robert Frost famously made this remark: “A jury consists of twelve persons chosen to decide who has the better lawyer.” Nowadays, this tongue-in-cheek line even applies to lawyering when it comes to the use of AI chat transcripts. Be aware and be prepared.
