Kasner estate-planning symposium turns AI use into a professional-responsibility question

Artificial intelligence has been promoted from productivity tool to professional-responsibility problem for trusts and estates lawyers, judging by WealthManagement.com's account of the 22nd Annual Jerry A. Kasner Estate Planning Symposium. The conference was held Aug. 27 and 28 and hosted by Santa Clara School of Law, and the publication's write-up says five key developments were highlighted; its account runs through three, stopping partway into the third.

The first is the one the account calls a central theme. Courts, regulators and bar associations are increasingly focused on how attorneys use AI rather than whether they use it, a through-line it traces across recent case law, proposed California legislation and State Bar guidance. The consistent message, as the account renders it, is that attorneys remain fully responsible for the accuracy, confidentiality and legal sufficiency of AI-assisted work product, and that transparency with a court is generally preferable to trying to hide the use. Practitioners may soon face affirmative obligations to verify cited authorities, disclose certain AI use and maintain appropriate oversight and documentation.

The advice aimed at firms is procedural and specific: establish clear AI policies, put citation-review procedures in place, track AI developments in local court rules and consider whether future pleadings, petitions, declarations and verifications will need to address AI-assisted drafting.

A second development moves AI into litigation. Presenters discussed the growing use of technology-assisted review and other machine-learning tools to find relevant documents inside large collections of emails, text messages, Teams chats, cloud-based communications and social media data. Those tools can cut review costs and improve efficiency, but a collection assembled with them still has to be defensible, which is why search terms, custodians and collection parameters should be documented carefully enough to survive a later discovery dispute. The drafts may not stay out of the record either: the account flags that AI-generated drafts, prompts and related communications may themselves become subjects of discovery, reinforcing the case for internal governance and recordkeeping.

The third item is where the write-up runs out. One of the more significant tax developments discussed at the conference, according to the account, was the proposition that an interest being difficult to value does not mean it cannot be gifted. The extract ends before the supporting discussion, leaving the fourth and fifth developments the conference highlighted undescribed in what is available.

That the conference addresses a law-firm audience does not confine the subject. The instruments at issue — entity interests that resist appraisal, trust documents, the email threads that produce them — live in client files, which suggests the governance questions aimed at law firms are ones wealthy families can reasonably put to their own advisers. The account identifies the proposed California legislation and State Bar guidance as where any coming obligations would originate, and attaches no deadline to either.