Recently, I received a Larke v Nugus request from the deceased’s second-born child (let’s call them  X).

For those unfamiliar with the term, a Larke v Nugus request is a request for information concerning the preparation and execution of a Will. It is commonly made where a beneficiary or family member has concerns about the validity of a Will and seeks information from the Will drafter before deciding whether to pursue a formal challenge.

The request did not come from a solicitor’s office but directly from the individual and was addressed to the executors.  X requested a copy of the deceased’s executed Will together with information concerning its preparation and execution.

Upon reviewing the request, I noted that it extended beyond the circumstances surrounding the Will itself. Among other matters, X required information regarding estate assets, liabilities, inheritance tax matters, solicitors’ fees and aspects of the wider administration of the estate.

I acknowledged the request and confirmed that I would respond within a reasonable timeframe with the information relevant to the Larke v Nugus request.

What followed, however, was unexpected.

Over the following days and weeks, the correspondence expanded considerably. Emails referred to solicitors’ duties under the SRA, executors’ fiduciary obligations, probate procedures, inheritance tax matters, potential conflicts of interest and requests for the executors to step down from their role. The tone also became increasingly assertive, with references to potential complaints, legal action and regulatory involvement.

At first, I considered this to be an understandable reaction from someone who may have felt disappointed or excluded. However, it gradually became apparent that something was not quite right.

In accordance with the executors’ instructions, I responded to the Larke v Nugus request and disclosed a copy of the Will. Although  X was satisfied to receive the Will itself, the wider requests for information were continued.

Following a review of the correspondence, it became apparent to me that AI tools had likely been used to generate much of the content. For a litigant-in-person, this is entirely understandable and – in many cases – perfectly reasonable.

However, the experience raised an important question: does AI-generated legal assistance always help individuals understand their legal position?

In this example,  X appeared to believe that every issue raised fell within the scope of the request. Yet the correspondence did not distinguish between information that was relevant to a Larke v Nugus request and information that fell outside its purpose.

This highlights a broader challenge with AI-generated legal content. AI tools generate responses based on the information and assumptions provided by the user. Depending on how a question is framed, the same issue can produce very different answers, each appearing equally persuasive and authoritative.

Where an individual has limited legal knowledge, there is a risk that AI-generated responses may reinforce misunderstanding rather than provide clarity. The technology may provide information, but it cannot always explain the legal boundaries of a request or exercise professional judgment in the way a legal adviser would.

In other words, the individual may ultimately become “lost in translation”.

This raises an increasingly relevant issue for private client practitioners: the growing use of AI-generated legal correspondence by beneficiaries and litigants-in-person.

While technology can undoubtedly improve access to information and encourage individuals to better understand legal terminology and procedures, it can also create challenges for practitioners and individuals alike. AI-generated correspondence can appear legally sophisticated and authoritative even where the underlying legal basis is weak, inaccurate or misunderstood.

In probate matters particularly, there is a real risk that AI-generated content may blur the distinction between a legitimate request and a broad fishing expedition into estate administration.

For practitioners, this creates an additional challenge: distinguishing between genuine legal concerns and requests that have expanded beyond their proper scope using technology that cannot yet exercise legal judgment.

AI is undoubtedly transforming access to legal information. However, access to information is not the same as understanding the law.

The lesson is a simple one. In the age of AI, confidence should not be mistaken for accuracy.