As readers of this blog know, I have on rare occasions consulted an AI tool with respect to some issues I have blogged. I tended to dip my toe in and then stay out for a while. I have decided to start dipping my toe in more frequently and deeper.
I have been inspired now by Robert Wright’s The God Test: Artificial Intelligence and Our Coming Cosmic Reckoning, Amazon here, I have started reading the book and am impressed so far. Fittingly, I asked an AI tool, Microsoft Copilot the following:
Please point me to the best reviews of Robert Wright's The God Test: Artificial Intelligence and Our Coming Cosmic Reckoning and summarize the key points of the reviews.
Readers of this blog entry can read Copilot’s response here.
I have been a Robert Wright fan for a long time—since reading his book titled The Evolution of God , Amazon here, when it came out in 2009. (As an aside, shortly after it came out, I taught a 4-week Sunday School class at South Main Baptist Church, Houston, TX on this book.) Judging from these two book-end titles, one can infer that he thinks a lot about God and, from the content of The God Test, about the continuity of evolution.
Suffice it to say that I now am convinced that I need to use AI more to understand its possibilities and its limitations. So, what better way to start the process than to ask an AI tool to assist me. I used ChatGPT Plus (“ChatGPT”) for this effort. I started with the following prompt:
Can you review the entries on my blog to learn my interests and writing style and then use the information on new potential blog entries discussing new cases involving my subject, tax procedure?
I then used some iterative prompts to fine-tune how I wanted to use this tool. Readers of this blog can review my prompts and ChatGPT responses here. There is a lot in the responses, but for now my key bullet-point takeaways are:
- One of my prompts was: “Please do a more extensive review of your 2025–2026 blog archives to develop a reliable picture of my interests and style for future use.” ChatGPT reviewed my Federal Tax Procedure Blog entries for 2025 and 2026 to give its conclusions as to my interests and style. ChatGPT’s responses were informative for me, although in a sense I have lived it.
- I then asked ChatGPT: “Will you please periodically review new federal court opinions and suggest to me the opinions that offer potential blog topics in my areas of interest?”’
- ChatGPT will provide suggested blog content on Friday of each week after it has reviewed new cases in my area of interest
I may make some more iterative prompts which will show up in the link above. But that is where I leave it for now. Here is how ChatGPT describes the task:
- I'll use the editorial profile developed from your 2025–2026 blog archives to select decisions that fit your interests, particularly:
- Statutory interpretation, Loper Bright, Chevron, and administrative law.
- Assessment, penalties, statutes of limitations, and procedural safeguards.
- Tax Court jurisdiction, tax litigation, and judicial review.
- IRS authority, Treasury regulations, and taxpayer remedies.
- Cases involving questionable judicial reasoning or broader doctrinal implications
What you'll receive
I'll recommend up to five particularly promising opinions, ranked by their potential as blog topics.
For each, I'll provide the case citation and opinion link, a short summary of the holding, why I think the decision warrants your attention, and suggested analytical angles or possible blog titles.
I'll also check your previous blog entries where feasible to avoid unnecessary duplication.
Most importantly, I will favor cases that present interesting legal questions over cases that are merely newsworthy. If a court reaches the correct result through questionable reasoning, that may be an especially good candidate for your blog.
The weekly review is now scheduled. You can also ask me at any time to prepare a full proposed entry about any recommended case, using your established writing style.
Finally, I offer a recent article on how to make AI more useful:
Bryan A. Garner, What are some good prompts when using AI for legal writing? Bryan Garner has some thoughts (ABA Journal 9/29/26), here (this link likely requires ABA membership). The following are key excerpts:
Better questions
The least useful prompt is: “Write this brief.”
Better prompts sound like questions from an exacting partner, sharp opponent or excellent clerk:
• State the strongest counterargument. Don’t invent facts or authority.
• Identify every place where this analysis assumes its conclusion.
• Does this memorandum state the governing rule before applying it?
• Propose a better sequence.
• Mark every factual assertion that needs record support.
• Find vague referents, shifting terminology, unexplained transitions, and overlong or imprecise sentences.
• Revise this passage three ways: one plainer, one shorter and one more formal.
• I need to cut 1,000 words from this brief. Give me a version with every cuttable word and sentence in boldface so I can decide.
• Read this as a judge with limited time and no desire to rescue my argument. What will you find unpersuasive?
Those prompts don’t evade thinking. They require it.
You must decide what question to ask, supply the relevant material, assess the answer, verify the authorities, and decide whether the proposed change improves the work. The machine can generate possibilities. It can’t assume responsibility for choosing among them.
* * * *
AI won’t replace lawyers.
Lawyers who use AI effectively will replace lawyers who don’t.
No comments:
Post a Comment
Comments are moderated. Jack Townsend will review and approve comments only to make sure the comments are appropriate. Although comments can be made anonymously, please identify yourself (either by real name or pseudonymn) so that, over a few comments, readers will be able to better judge whether to read the comments and respond to the comments.