Sometimes I worry that AI is going to replace lawyers. And then the universe hands me an opinion like Llewellyn v. Fechtel, showing that nothing that able to pass the Turing test could ever do our jobs.

Llewellyn is a trusts-and-estates case; basically, Fechtel claimed the Llewellyn breached her fiduciary duties as trustee. The trial court granted partial summary judgment against Llewellyn on two issues. It got flipped on each, with Judge Lorish writing for a panel that included Judges O’Brien and Ortiz. (As an aside, this is a very fun panel. Credit to the algorithm for putting it together.)

First, the trial court relied on Llewellyn’s interrogatory responses to conclude that she’d admitted to borrowing about $75K of trust assets from a personal bank account. The Court of Appeals reversed, pointing out that interrogatory answers are not binding admissions and must be read together with a party’s explanations. Because Llewellyn maintained that the personal joint account was not necessarily funded with trust money, the Court found that genuine factual disputes remained and that summary judgment on this first issue was unwarranted.

Second–and much more hilariously–the trial court deemed thousands of Llewellyn’s responses to requests for admission as admissions that she’d misused another $235,346.35 from the trust. Except the RFAs had nothin to do with the authenticity of particular documents and really just boosted the plaintiff’s theory of the case. For context, Fechtel served 8,984 RFA spanning thousands of pages, of which 8,920 addressed 446 identified transactions. (Why even try to do this?) CAV concluded that the trial court had erred by failing to recognize the limits set by Rule 4:11(e), which caps RFAs at 30, with an exception for verifying the genuineness of documents. Most of Fechtel’s 8,984 requests sought facts and legal conclusions about the source and legitimacy of transactions, not the authenticity of any single document. To illustrate, the Court quoted RFA 337 (!) in its entirety, including all 20 subparts (!!). (Again, WHY EVEN TRY TO DO THIS?!) So the Court held that the vast majority of those requests exceeded the permissible scope and could not be used to support summary judgment.

Terrific opinion, with a helpful treatment of summary-judgment and discovery rules–all of which work only when courts enforce them. That’s actually the answer to “Why even try to do this,” right? Because maybe the trial court will let you get away with it. (And 5 years ago, maybe you evade discretionary appellate review.) Glad to see CAV shut this abusive silliness down.