Frankenstein’s Monster*

My March column, "The Horror, the Horror," prompted a masterly written response by Judge Michael Kelley. See Daily Journal column "Why we write" (Apr. 24). The title of my March column speaks to my concern about what may be an eventual, if not present, reality: ChatGPT writing judicial opinions. I borrowed the words spoken by the character Kurtz at the end of Joseph Conrad's novel Heart of Darkness to illustrate the horror of such a prospect. In a different time and context, Kurtz's words reflect his recognition that humans, himself included, commit acts that are evil. I see a possible parallel.

Judge Kelley's informative and engrossing column rejects this cataclysmic view and seeks to put me and others at ease. For now, I agree with his conclusion that "AI programs will never come close to producing an opinion as elegant and enduring as [Justice Cardozo's] Hynes [v. New York Cent. R. Co. (1921) 231 N.Y.229]; and why they should never be asked to try." I leave to you, dear reader, to read this and other Cardozo opinions, most of which shine with clarity and elevate prose to the height of elegance.

But Judge Kelley's superb column does not entirely convince me to rest easy. In fact, it prompts the title here, "Frankenstein's Monster*." Does anyone write judicial opinions with the clarity of expression and choice of just the right words that come close to matching Justice Cardozo, Justice Holmes, and I will add our own Chief Justice Traynor?

Yes, AI programs do not replicate Shakespeare in subtly, wit and irony. Nor do they come close to matching Cardozo or Holmes in a judicial opinion. Judge Kelley put an AI program to the test and asked it to provide a description of proximate cause in the style of Cardozo. Here is what it delivered: "To approach the legal doctrine of proximate cause is to grapple with the quintessential principle in the law of torts. Its essence lies in determining whether a defendant's actions are causally linked to a plaintiff's harm. Proximate cause signifies the legal notion that a defendant can be held accountable solely for the harm that was directly caused by their actions and that was reasonably foreseeable."

Hmm, reminds me of many judicial opinions I read today. A.I. can write judicial opinions. They do not come close to matching those of Cardozo. I'm hard-pressed to find any that do. There is a suggestion afoot that many judicial opinions today are verbose, vague, and at times incomprehensible. Who am I to disagree with myself? My reference is to opinions throughout the United States. To avoid my moniker "troublemaker," in this column, I did not limit my critique to California. I acknowledge that poorly drafted recent legislation has contributed to some confusing California judicial opinions. I include my own efforts in the mix. To avoid melancholy, I avoid reading those I wrote in the past. But the prospect that A.I. could be used to write judicial opinions, however lacking in graceful expression, engenders in me a more pervasive melancholy.

But as an aside, I offer this caveat: even the masters can miss the boat. Cardozo's opinions shine with clarity and poetic prose, but the legal discussion in his famous Palsgraf v. Long Island R. Co (1928) 248 N.Y. 339 is another story. Here I change from the metaphor boats to a real train. Who can forget the description of the two men running to board the moving train, the package of fireworks falling, the explosion causing the scales to fall on poor Ms. Palsgraf? The legal reasoning to deny relief to Mrs. Palsgraf could have been written by an AI program. (Bracing for the pushback.) Before any of you throw the first stone at me, please read Justice Andrew's dissent. I will save this discussion for another column.

Getting back to Judge Kelley's column, please note there is an asterisk after the title of this column. Judge Kelley summarizes the thoughts of some professors and jurists about judicial opinions as literature. His takeaway? "A rationality that is capable of 'literary imagining and sympathy'" does not spring from an algorithm. It must come from the intellect and lived experience of a human judge." This frame of mind leads to our triumph over A.I. But it requires an unwavering commitment and a conscious effort. In other words, more than a draft or two.

We may not match Cardozo in felicity of style, but we can be clear and concise. Our job is to communicate a reason for our decisions. We must write to enable readers to understand what they can and cannot do. We must do so with clarity and concision. We must write in a manner that holds the reader's attention. Forty pages of rambling defeats this goal and allows A.I. to intrude where it does not belong.

The manner of expression, the unique arrangement of words that communicate the reason for a decision, is what we humans do better than a technology that in milliseconds searches and spits out words that humans used in the past. Familiarity with literature and the humanities as a human, as opposed to an unthinking machine searching for words, is our triumph over A.I. It may not come to an end like Frankenstein's Monster, but we can and must keep it under control.