Use of AI in Family and Probate Law
Full disclosure: to illustrate for the reader how AI is being integrated into the practice of law, this post was created via collaboration between the human author [Timothy Flynn] and Google's AI tool. The approximate word count break down is 75% human authorship and 25% AI generated content. The AI portion is noted in the text below. The image contained in this post [left] was generated by Google's Gemini platform.
Advancement of AI has polarized us into two primary groups: those who believe that AI is a savior for humanity and those who believe AI will bring about the demise of humanity; think Skynet from the Terminator franchise.
Although AI has taken the world by storm over the past 5-years, large language model AI has been instrumental in legal research since the 1980s. This post goes beyond the legal research capabilities of AI and examines how AI has affected the legal industry as a whole, particularly family and probate law practices.
In the early days of computerized legal research, practitioners would input complex Bolean search requests and the computer would output a list of relevant casees and statutory cites. Then the attorney would manually deconstruct and assess the output, utilizing what was useful or on point.
These days, everyone with access to the Internet has access to a broad menu of AI: Claude for in-depth detailed research; DeepSeek for reasoning and logic; Gemini for generation of images; and Chat GPT for light-hearted Q and A.
Westlaw is Making AI Available to Judges
When our law firm investigated Westlaw's AI product, we learned about some amazing developments that are already out of beta and now on-line. One AI leagl research product is available only to judges. The software scans a Plaintiff's brief, scans a Defendant's response to the brief, analyzes all the caselaw and statutes cited in the attorney work product, then prepares a "recommendation" for the jurist.
On average, this process removes hours and hours of time from a judge and her staff. The bottom line, however, is that even though a tough leagl issue is distilled and handed to the jurist on a platter, a tough decision still needs to be made. What's more, in close cases, it is the very process of going throug the details of the filings that will inform the practitioner and jurist alike as to the conclusions that should be reached.
Admittedly, judges have a more difficult task to the extent that they are not paid to advocate; they are paid to make a reasoned decision based on the application of the law to the facts of each case. We here at Clarkston Legal are quite certain in such cases, we want the jurist making those decisions -even if advers to our client- rather than a machine.
The temptation to let the machine do the heavy lifting is powerful. In the end, however, it is our belief that the judge needs to make a well-reasoned independent -human- decision.
Classic Legal AI Missteps
Earlier this month, AI struck again, when pillow mogul Mike Lindell's lawyers filed a brief in federal court in Colorado containing ficticious case citations; Lindell's lawyers were sanctioned $3000 by the judge. No one knows why AI is unpredictably prone to occassional "halicinations" but the practice is highly unsettling, especially if you are involved with high-stakes litigation and you are depending on lawyers to file briefs with federal courts on your behalf.
This type of halliucination also occurred in a New York case where opposing counsel could not find the cases sited in the opponent's brief. This was because the AI tool used to write the brief hallucinated to fictitious caselaw. The lawyer that propounded the brief did not proof the work product output of the computer. Never a good move. Here is the link to this earlier post with our analysis.
Some federal judges, perhaps in reaction to these ridiculous AI missteps and case site halucinations, require certification by the filing attorney that AI did not play any part in drafting a filed pleading, brief, or other document. If an attorney violates this certification, the attorney is subject to sanction by the court.
Family Law and AI
When pundits warn about AI replacing humans in various industries, they are not referring to lawyers that appear before referees or judges in family court. Fortunately, with each case depending on its own circumstances and depending on the local personalities involved in the local county family court, AI is a long way off from even supplementing the work load, let alone taking over.
Bottom line in family court is that the personalities need to be present to carry the day. This includes the parties themselves, their lawyers, referees and of course the judge.
Of course, complex property issues and extensive legal briefing are always subject to an AI assist. We have high-net-worth individuals where AI generated briefing and analysis comes into play. Also, our law firm's new client management software has a useful AI component that allows us to bring that functionality into the courtroom.
Most often, however, our lawyers walk into court sans computer and go toe to toe with opposing counsel over the facts and relative equities of a case; old school. Overall, family law is just not cut out for a significant AI assist.
Probate Law and AI
Probate law tends to be, on average, a bit more document-intensive. Therefore, probate law lends itself to AGI assistance more readily than family law. For example, new AI programs are available that allow lawyers to review their myriad extensive trust and estate planning documents for changes in the law or any number of custom filters or objectives.
Trust documents are not "one-size-fits-all" to be sure. At our law firm, trust instruments are customized to the clients' needs and fit securely into a comprehensive estate plan. Estate plans are customized based on the circumstances and the intent of the settlor.
Wealthier individuals often have tax issues embedded into their estate planning needs. AI legal software is now available that can crunch the time it takes to review, analyze, and integrate complex tax provisions into a trust instrument. The software can then ensure that all the other component documents within a particular estate plan are consistent with the tax treatment of various trust issues.
Probate litigation is again bringing the disputes -often family disputes- into the courtroom of a local probate judge; usually with local probate practitioners. Depending on the crux of the dispute, AI may or may not be very useful in resolving the dispute.
State Bar of Michigan Addresses the Use of AI in the Practice of Law
Now we reach the portion of this post where AI provides a useful service in the creation of a law blog post. The State Bar of Michigan's June 2025 publication, "Transforming the Legal Landscape in the Ago of AI" is 56 pages long. A useful AI query is to have the document summarized. The key components of the report are summarized in the italicized paragraphs below and are not the origianal work product of the author.
The document is a report on the integration of generative artificial intelligence into the legal profession, exploring its opportunities, challenges, ethical considerations, and implications for practice management and access to justice.
The key areas of AI's impact on legal practice, as outlined in the document, include:
Document Review & eDiscovery: Automating the review of electronically stored information (ESI) to save time, reduce costs, and improve accuracy.
Legal Research: Streamlining research processes using natural language searches and large language models (LLMs), while addressing risks like hallucinated responses.
Legal Drafting: Generating drafts for legal documents, contracts, and briefs, enhancing efficiency but requiring careful review for accuracy and ethical compliance.
Contract Drafting & Analysis: Assisting in drafting and analyzing contracts, identifying key provisions, and reducing human error.
Document Storage: Managing electronic storage with considerations for security, retention policies, and compliance with record-keeping laws.
Outcome Prediction: Using AI tools for case prediction, jury analysis, and judicial rulings, though limited by human behavior and biases.
Cybersecurity: Integrating AI into security frameworks to analyze data, recognize patterns, and adapt to threats, while ensuring client confidentiality and data protection.
These areas highlight AI's potential to enhance efficiency, reduce costs, and improve legal services, while requiring vigilance to address ethical, accuracy, and security concerns.
The risks of unauthorized practice of law (UPL) with AI include:
Blurred Line Between Legal Information and Legal Advice: AI tools may unintentionally cross the line from providing general legal information to offering legal advice, which requires legal discretion and profound legal knowledge.
Lack of Regulation: AI-based legal services are largely unregulated, raising concerns about the accuracy and reliability of the information provided.
Potential Public Harm: Incorrect or misleading legal information generated by AI could harm users who rely on it without consulting a licensed attorney.
Scrivener’s Exception: AI tools generating legal documents may fall under the scrivener’s exception, but this depends on whether legal knowledge or discretion is involved, creating ambiguity in enforcement. The scrivener's exception to unauthorized practice of law (UPL) allows non-lawyers to assist individuals in completing standardized legal forms without selecting the form or providing legal advice on what information to include. This exception applies when no legal knowledge or discretion is involved in the document's completion.
For AI-based programs, the scrivener's exception raises questions about whether automated form generation falls within this exception, especially as AI tools become more advanced and widely available. The accuracy and reliability of AI-generated documents will be critical in determining whether they comply with this exception or risk violating UPL regulations.
Challenges in Enforcement: Michigan’s UPL statute applies to "persons," making it unclear whether AI programs or companies using AI can be held accountable under current laws.
Bias and Hallucinations: AI tools may produce biased or fabricated outputs, leading to flawed legal advice or documents.
Public Misunderstanding: Consumers may struggle to distinguish between legitimate AI legal tools and substandard or deceptive services, increasing the risk of exploitation.
These risks highlight the need for careful monitoring, public education, and potential legislative updates to address the evolving role of AI in legal services.
Ok, back to human authorship. One area that was not addressed in the SBM's report was: client privacy within the context of AI assisted research. Perhaps this concern was deemed outside the scope of the SBM's task force. It would be intuitive and typical for an attorney to input various details -some of them identifying- into a large language model AI platform in order to get a desired output. The more detailed and refined the inputs, the more accurate the output.
Some clients may not want the details of their case to be floating around in an AI platform, potentially accessible to third parties. To combat this, some of the platforms like Chat GPT have rolled out a "law" version of their service -ChatGPT 4- that guarantees confidentiality of both query and output. The monthly fee for the enhanced service is approximately $60.
AI is Coming to a Law Firm Near You
The march of technology is relentless; we cannot avoid it, but we can temper its impact. One little known fact is that the recently passed so-called "Big Beautiful Bill" contains a decade long ban on the state regulation of AI. The congressional act proscribes nearly all forms of state laws designed to regulate or limit AI platforms.
So take note state legislatures, Congress has declared a "hands off" approach to the development of AI over the next decade. That's a long runway for the industry to stretch its long legs. Here is a link to a very "optimistic" assessment of where AI technology will be prior to our next presidential election; it's called "AI 2027".
We here at Clarkston Legal have always taken a human client-focused approach to the practice of law. We are not going to change that focus while machines gain ground in deliverying efficiencies within the practice of law.
The legal process is at base a human process. Family law and probate disputes are human disputes. Local family and probate jurists, at least for the foreseeable future, are human jurists. So within the context of technological advancement, we will continue doing what we do best: delivering excellent legal services at a reasonable fee to our human clients.
Post #308