To Use AI Slop, Or Not To Use AI Slop

We all know it when we see it. AI slop. Articles and documents written in whole or in part by AI. The one-sentence paragraphs. The “It’s not X. It’s Y.” The solemn proclamation that somebody is saying the quiet part out loud. The use of the em-dash. The kicker. 

Ugh. Many of us hate it. But we need to steel ourselves. As AI content proliferates, the slop may inevitably become not only accepted but the expected way to communicate. 

Sounds like a nightmare. But remember styles of writing have always evolved. The language of Shakespeare’s plays, while beautiful, is not the style in which we typically communicate today. Nobody would send an email today: “Hark, Steve from law firm, what evil tidings dost thou have today regarding my lawsuit.”

But the transition from what is expected today to an AI style could have enormous impacts on lawyers who are in the business of communicating and persuading. Think it won’t happen? A recent set of studies suggests that transition has already begun.

 The Slop May Be Winning

The AI slop haters (which, as you know by now, includes me) all think that just because we hate slop, everyone else does as well. That henceforth and ever more AI slop will be roundly rejected, ostracized, and criticized. But it’s always good to look at some data before we continue to be so self-righteous. And there is some.

The Cambridge University Press recently published the results of some studies that tested whether people could tell the difference between human writing and AI writing. And to see which they liked better. Over 1,600 participants read either a human written story or a similar AI written story. Half were told the story was written by a human, and half were told the stories were written by AI. Half were told the truth about the author; the other half were not.  

The first finding should be comforting for those of us that despise AI slop: the participants who incorrectly thought they were reading human stories gave those stories higher scores than the ones who thought they read the AI stories. Clearly, there is a bias against what is thought to be AI writing.

Unfortunately, there’s more. The real headline is the conclusion that the participants, in fact, liked the real AI-written stories better. They thought the AI stories were more engaging and of better quality. That’s saying the quiet part out loud for you.

The reasons, according to the study:

AI-generated content is often more fluent and easier to parse than human-written content, and people have a well-known tendency to prefer material with those characteristics. Additionally, AI-generated content tends to have a more positive emotional tone, which people also perceive more positively.

That’s the same thing as saying writings in the style of Shakespeare are harder to read. 

And speaking of Shakespeare, the study goes on to cite other work that paired poems by Shakespeare with similar poems generated by ChatGPT. The participants in that study somewhat ironically thought the true Shakespeare poems were more likely to have been generated by AI. 

Think about that: the most celebrated writer of the English language was thought to be a bot. The point is the participants thought that poems written in a style with which they were unfamiliar were not as good as those written in the style they were. (The Cambridge University Press study contains documented support for all these conclusions, by the way.)

So we may already be at the point of no return when it comes to AI slop.

So What for Lawyers

It’s axiomatic that as lawyers, we have to communicate in ways that our audience expects and best understands. Just like those in the Shakespeare poem study, audiences like a style with which they are familiar. And that is the style more likely to persuade. It’s the old meet people where they are notion. 

So if the Cambridge study’s conclusions are correct, that means, more and more, we are going to have to communicate in, God forbid, the style of AI. Otherwise, we risk the audience getting distracted and bored, and the likelihood of persuading them decreases. It’s no different from using and trying to manipulate poster boards in a courtroom as opposed to screens. The latter is generally more effective today than the former. The type of information, how we process it, and how we learn is constantly changing, and we have to change with it.

The Business We Are In

I know. Many are recoiling at the prospect of having to try to sound like a bot. But all is not lost. Shakespeare’s writings have stood the test of time and are revered not simply because of the writing style but because of the lessons they taught. What they said about what it means to be human. His writings taught us things about ambition, jealousy, fear, love, betrayal. Shakespeare’s plays may have been written in the style of the day, but that doesn’t detract from the impact the plays still have.

The same is true of the lawyer’s job. We are in the business of communicating and persuading. Communicating and persuading is not all about style of writing and speaking. It’s understanding what points will be the most persuasive. What facts matter. What examples and analogies resonate the best. It’s about how we frame and define the issues. Most importantly, it’s how we tell the story most persuasively. 

Think about the Cambridge study. The experience and storyline were created by a human. That storyline was fed to AI, and it created the language by which the story was told. The underlying story and the experience were still human ones.

We will still have to make sure, whatever the style in which the key points are communicated and the message, the story we want to tell is conveyed. We will still have to be sure the AI slop style conveys the message and story most effectively and convincingly. 

The danger, of course, is that we don’t separate the style from the message and believe AI can do the best job at both, as tempting as that may be. Turning the messaging and storytelling over to an AI bot means we aren’t going to be as persuasive as we can be. Not using AI style may soon also detract from our persuasive effectiveness.

But There Is One More Thing

We aren’t yet at the point where AI writing is the norm and therefore consistently the expected. But we’re headed that way. That places the lawyer in a difficult spot. It may very well be that some of the lawyer’s audience lives in the AI world and expects the AI communication style. Some of the audience may be like me and recoil at anything that even remotely resembles something AI-generated.

As lawyers, we may have to accept the fact that we will now need to communicate using both styles. It’s the same reason lawyers tell their story using different communication tools. Some people learn through video, some through paper. Lawyers often use both. I often felt the need to use both in a trial even at the risk of being repetitive to reach as many jurors as possible. 

And if the Cambridge University Press study is correct, our audience may already be transitioning to more acceptance of AI-generated content. I would be the first to admit I don’t like where we are and, more importantly, where we are headed. AI writing is mostly slop, and to me it detracts from whatever message is trying to be conveyed. But it’s going to become more the norm, and we better accept it.

But for those of us who still gag at the sentence that starts, “here’s the thing”: take heart. It’s not whether to use AI slop or not to use AI slop. At the end of the day, it’s all about the message and story. The style may need to be AI slop; the most effective message and story won’t be. 

---

Stephen Embry is a lawyer, speaker, blogger, and writer. He publishes TechLaw Crossroads, a blog devoted to the examination of the tension between technology, the law, and the practice of law.