AI can make you lose your powerful Attorney-Client privilege

The difference between revealing your inner secrets and genuine concerns regarding your legal matter to a friend and sharing your legal problems with AI is that you already know that your friend has somewhat limited knowledge of the legal options available to you as it relates to your legal matter and can provide limited feedback; therefore, you can try to find answers to your legal problem only to some extent, but you do know that AI does not have these limitations. AI is expansive in legal knowledge and thus gives one the impression that it is unlimited in the options, questions, thoughts, or feedback it provides to the seeker. Therefore, you feel emboldened to open yourself up to AI and provide significantly more and more details as you read through the AI’s answers and follow up with other questions. In essence, you end up sharing with a machine not just your facts but also your own doubts, concerns, fears of perceived or actual liability that you have regarding the subject you are inquiring about and, further, dissecting those facts, doubts, concerns, etc. to exhaustion.

On the one hand, AI is indeed a technological advancement in our society for machines to expand human reach so that we are more efficient. On the other hand, however, you end up in a disastrous position to 1. have waived your attorney-client privilege; and, 2. help the opposing side and counsel with little effort on their part to prove your fault. Because what you shared with the AI machine is not protected by attorney-client privilege, the opposing counsel can have access to everything you shared with the AI and will have no hesitation in using the AI default to prove his case using your own statements against you.

The AI is here to stay. However scary it is right now, it will and does help our lives tremendously in so many ways. We will do more in less time and will accomplish things that we cannot do so easily currently. In short, we will not be able to perceive our existence without it.

Some say that AI scare is internet scare. You know, many “centuries” ago (at least it seems that way), people feared the internet. The internet was perceived as a monstrous technology because it provided a platform for anyone to post content for everyone else to see, so there was a sense of personal violation and people felt that they had no control over it. Further, the ease and speed with which information and knowledge was ready to be shared felt uncomfortable to many, particularly the technology-averse segment of the population in general.

The rise of AI is simply a natural development of the internet but is updated with the ability to retrieve information with the speed that our current lives crave – so, it is faster and produces more. Information can be widely distributed. Give AI a prompt and it can create tens of thousands of articles in a fraction of the time compared to a human. Most importantly, it is not reserved to the few that can afford it but instead offers unfettered access regardless of one’s financial ability. Society thrives on knowledge and avenues to share it should always be encouraged and supported.

There are indeed some drawbacks to the rise of AI and they are present across the spectrum. A student mistakenly believes that he/she can get away with having his/her essay produced with no or little effort on his own, only to then face the brutal reality of shame and no credit for his or her effort. A hacker can now be more efficient in breaking through the previously erected technological barriers and penetrate information to create damage to someone else, financial- or reputation-wise.

In the legal world, there are, however, dangers of AI that are less known, primarily because not everyone has a legal problem to apply this tool toward to and understand the ramifications of its use. Furthermore, even though AI application to legal questions can have consequences for both lawyers and clients when the legal citations produced by AI are inaccurate, these consequences tend to be more severe for clients, particularly if they have not had to previously deal with attorneys and do not understand the protection offered by the attorney-client privilege umbrella. [1]

When people are facing challenges, particularly legal, they feel truly helpless. Not just because of the challenge itself and the disruption in their lives, but also because of their naivete as it relates to what is their next step. And it is natural to not know what to do amid chaos – even if the legal matter that you are dealing with is at a lower level of concern and more predictable, such as forming a company. If you are not a lawyer, and even if you are but do not practice the area of law associated with your matter, you feel lost and apprehensive as to how you will accomplish what you need or protect yourself or assets in a dispute. I would like to add that even though in most litigation matters one has some awareness of the potential legal trouble brewing on the horizon, many clients are still in shock when they are hit with a legal matter, scrambling to see through what they should do next.

The prospect of seeking legal help is in and of itself a traumatic experience. First, clients know from what they read or their friends’ own experiences in the legal realm that the legal fees can be quite exorbitant and, naturally, spending hard-earned savings on legal bills is not a thought bound to alleviate their current angst. Second, the difficulty in finding an attorney that clients are looking for is quite a journey that usually consists of long days and long nights of browsing the internet for websites to find someone who can come to the rescue. The ideal lawyer that a client is looking for must demonstrate a trifecta of qualities: ability in delivering an immediate pathway structure to achieve the client’s goal (which entails knowledge), leveling with the client (which also entails empathy), and personality match, the last factor possibly outweighing sometimes the other two. Third, once they finally find a lawyer that manifests those qualities, at a reasonable level, then they will have to face the next barrier, which is whether they can afford that lawyer. The arduous efforts expanded up to this point will have clearly visible effects on the person’s entire being, physically, emotionally, and psychologically.

As you can see, finding legal help is an ordeal in and of itself which only exacerbates the current psychological and stressful state that the client is in. Therefore, it sounds only natural that people will rush to use any tool at their disposal that could give them some relief from their current turmoil. With the rise of AI, such a tool seems not only readily available but also effective in providing the person who is seeking it with some knowledge, at least at the basic level, to alleviate their concerns. “Seems” is too weak of a word to describe the actual effectiveness of using AI, as it is currently available, in researching answers to legal inquiries, as the accuracy of its results has been questioned extensively by courts.

First, it is known that AI can “hallucinate,” in other words, produce research results and provide answers that are persuasive in substance, yet are based on statutes or case law that do not exist. Attorneys themselves have fallen prey to using AI for purposes of finding case law and citing the same in their court briefs, only to be admonished by judges whose clerks were able to quickly identify the absence of such valid legal theories. It is not only the embarrassment of acknowledging, without defense, that they chose to use a still unknown, to some extent, tool to do their research, but also that their arguments based on such results were flawed, a weakness that will be gleefully exploited by the opposing counsel, of course, should the opposing counsel be able to present viable and solid arguments that are based on real and valid legal sources. It follows, therefore, that if an attorney cannot distinguish the veracity of the work generated by AI, it is reasonable to assume that it is theoretically impossible for a lay person with little or no legal background to do so. [2]

The harm that AI may cause could be irreversible. The impact is abrupt and dual in that AI may provide information that is flawed, and therefore you may be planning a course of action that is unlikely to manifest in practice or obtain the outcome that you are looking for; additionally, AI can expose the client to an unwanted waiver of his attorney-client privilege.

The most severe impact the AI has on people looking to it as the immediate hero of knowledge coming to the rescue lands on the essence of the attorney-client privilege.

Among all privileges, the attorney-client privilege is one of the most sacred privileges, along with spousal, clergy, and doctor, as the primary ones involving two parties; all privileges are instrumental in maintaining the freedom of speaking and sharing openly your innermost fears and concerns. Yet, the attorney-client privilege is most ignored; maybe it is so because the significance of its benefits of using it and the dangers of losing it are not well known by the public.

Simply put, you say to your friend: “I am in big trouble. I made a mistake at work that is costing my boss millions of dollars in losses.” Your friend will lend you a shoulder to cry on and an ear to listen. If events evolved to such an extent that you are now in an active civil litigation dispute with your boss, we are now assuming “former” boss, whatever you said to your dear friend it will now be revealed through a legal procedure called “discovery.” This means that since your statements as to what has occurred and your own perspective, such as your admission or negation of fault, are not made to an attorney but to a friend, and since there is no “friend privilege,” the opposing counsel can compel your friend to restate everything you shared with. For purposes of our article, your “friend” is AI with whom you converse at a much higher risk of damage to your defense, if you are a defendant, or the possibility of an award or the size of it, if you are the plaintiff.

This naivety is not to be blamed entirely, or maybe to any extent, on the client, as the concept of attorney-client privilege is not well understood by many, including some lawyers. What can be attributed, however, to the client is their choice. Choosing AI is easy and free. However, opting for something easy for resolving something difficult without considering the ramifications is reckless. Believing that the information that you share with AI on a server stationed who knows where (it is not in your attic) cannot even be discovered, in the true sense of the word, is naïve indeed and a pure judgmental mistake. Clients may believe that skipping a lawyer to receive an answer on a legal matter will save them money; this attitude is not something to be blamed – we all know legal bills can be quite high and no one likes to pay them. Tell me that you know someone who cannot wait to pay a lawyer – there is no one. However, you must make wise choices and weigh on whom you confide in and what this could mean for your matter, considering a worst-case scenario.

It can still get worse. You feel so comfortable with AI that you may find yourself inclined to share what your own lawyer told you, particularly when you don’t like the answer. Sharing the conversation you had with your lawyer with AI may make it eligible for discovery, and, in this case, you may waive the attorney-client privilege not only as to what information you share or questions you ask AI, but you are now exposing to discovery the information that your own attorney shared with you. This means that the opposing counsel can easily request that all conversations with your lawyer be turned over for his delight in winning the case against you.

Usually, shortcuts are bound to miss the target, which in this case is the resolve of your legal matter. AI can help you get there – I mean miss the target. This is, of course, a sarcastic comment, not because I devalue its significance but rather primarily because I find it extremely alarming and want to shout it out for everyone to hear: be careful what you ask AI and what information you provide in exchange for the answer.

Knowing this, we do want AI. AI has the potential to help us evolve as species in knowledge at a higher speed with which we access such information. Please be at peace with this thought: AI does not and will never replace humans. However, we must undertake efforts to inform ourselves on how to use AI for our benefit instead of our detriment.

For more discussions about lawyers, clients, legal practice, and the realities of the legal profession, visit AskSimina.

Cheers,

AskSimina

*The content above is simply an opinion that could be faulty and not thoroughly researched and should not be taken as an evidentiary fact, advice, or rule to follow. This is not legal advice. Please consult an attorney for your legal question or matter.

External links added

1. AI and attorney-client privilege / confidentiality: The ABA’s Formal Opinion 512 specifically addresses lawyers’ duties concerning competence, confidentiality, client information, and the use of generative AI. ABA Formal Opinion 512 — Generative Artificial Intelligence Tools
2. AI-generated legal citations and hallucinations: The ABA discusses the risks lawyers face when relying on AI-generated legal work and emphasizes that lawyers remain responsible for the work product they submit. ABA — AI and Attorney-Client Privilege: A Brave New World for Lawyers
3. Attorney-client privilege and waiver: Federal Rule of Evidence 502 provides important context concerning disclosure and waiver of attorney-client privilege and work-product protection. It also shows why statements about waiver should be treated carefully because waiver can depend on the circumstances. Federal Rule of Evidence 502 — Attorney-Client Privilege and Work Product