From Family Trees to Qualified Immunity: Reflections at 200 Weeks

This post marks 200 consecutive weekly posts in my streak, or three years and 9 months. I’m not bragging about the length of the streak. But I am pretty proud of myself.

Genealogy Finds

I got an email alert from Family Search Wednesday that said, “You have 2 new memories of your relatives.” Okay, cool. I clicked through to find what was essentially an ahnentafel for Johann Justus Henckel Sr., supposedly my 6^th great-grandfather.

I didn’t have him listed in any of my records.

He’s part of the Judy/Tschudi line, which my mother descends from. My dad only got as far as Amos Judy, who would be Johann Henckel’s great-grandson. This find, assuming everything tracks (and it seems to), adds over 50 people to my family tree, and that’s just the ones leading up to Johann Henckel. His tree goes at least a couple of generations further back.

AI “Writing”

Someone at AbsoluteWrite shared this article at the NY Times: “The New Fabio Is Claude” or, “Can AI Chatbots Write Emotionally Rich Romance Books?” (Gift link or archive link) It talks about how one romance author has generated “dozens of novels” spread over 21 pen names using Anthropic’s Claude.

I find it ridiculously ironic that an author is using an LLM client that was trained on stolen writing.

At any rate, my comment on the thread at AW was “I will confess that out of morbid curiosity, I’d find it interesting to watch the entire process of someone generating a novel using AI. I feel like the prompt would end up being as long as a decent synopsis.”

Sure, Claude or ChatGPT or Grok can generate a 65,000-word novel in “seconds to minutes,” according to ChatGPT. But as the software points out, once you factor in chapter generation, waiting on user prompts or feedback, and editing options like revising tone, pacing, and so forth, it can take hours or days to complete the process.

I’ve played with ChatGPT a few times beyond the headline and social media posts I currently use it for, and the main thing I’ve figured out is that the more input you can give it for a particular task, the better. So if I want the software to generate a novel, I need to give it as much detail about the setting, the characters, and the plot as I can. I can’t just give it the names of my characters. I need to describe them in great detail, both emotionally and physically.

But the NYT article points out that the chat clients can fixate on phrases or character details that end up costing you extra editing time. Note that the “ragged prayer” example occurred for multiple authors. I’d guess that’s because of the various works the large language models were trained on. (And what even is a “ragged prayer?”)

At that point, what have I gained by using the LLM client?

And even then, after all these hours of writing prompts and editing the resulting text, the person running this process generally can’t copyright the generated text, at least not in the US. International law could, at best, be called varied.

Qualified Immunity

Qualified immunity is a legal principle (some say legal fiction) of federal law that protects government officials from civil damages when they perform certain optional functions. To get past QI, the plaintiff has to show that a reasonable person would have known that the action the government official took was illegal. It most often comes these days when a law enforcement officer is accused of violating someone’s civil rights, and it usually involves the use of force.

It amounts to the officers saying, “I didn’t know that was illegal.”

But ignorantia juris non excusat, right?

This doctrine, invented by the Court out of whole cloth, immunizes public officials even when they commit legal misconduct unless they violated ‘clearly established law’. That standard is incredibly difficult for civil rights plaintiffs to overcome because the courts have required not just a clear legal rule, but a prior case on the books with functionally identical facts.

Jay Schweikert, writing for the Cato Institute’s Project on Criminal Justice and quoted in this Forbes article.

That last phrase is what trips people up when they sue cops.

“Functionally identical facts” can be disturbingly specific. It’s often not enough to pull up a case where an officer used unreasonable force against someone in a wheelchair. A plaintiff might have to find a case where a white rookie officer used a Taser against an Asian woman in a wheelchair.

Here are more examples of qualified immunity cases at Reason. The facts in each case vary widely, and sometimes it’s stunning that officers receive QI.

The Institute for Justice, through its Project on Immunity and Accountability, founded Americans Against Qualified Immunity, a grassroots initiative devoted to fixing the QI mess. Learn more about AAQI, IJ, and the Project here: https://aaqi.org/about/

Thanks for reading. Feel free to share a thought in the comments. Sign up for my infrequent newsletter here. Find some of my other writing at The Good Men Project, too. Subscribe to the blog via the link in the right sidebar or follow it on Mastodon. You can also add my RSS feed to your favorite reader.