A.I. litmus tests are counterproductive

Over the weekend on Threads, the social media site where writers go to experience collective freak-outs, the main character of the day was a literary agent who, in short order, both complained about A.I. slop in query letters and said A.I. can be useful for researching comp titles.

I don’t want to link directly to the ensuing uproar lest I unfairly single anyone out, but my synthesis of the tidal wave of responses:

How DARE an agent ADMIT TO USING A.I. which is an EXISTENTIAL RISK TO EVERYTHING HOLY.

I would NEVER want an agent who USES A.I.

THE ONLY ACCEPTABLE A.I. USE IS NO A.I. USE

On Threads, not exactly the place for nuance, it apparently can’t be true simultaneously that 1) A.I. slop writing is bad, and 2) A.I. can be useful sometimes, particularly for things like comp titles.

I’m sure we can all agree that slop is bad. I’m not defending slop. But good luck policing A.I. use among authors (and agents, for that matter).

I may be sticking my finger in the light socket today, but I must admit that I believe the stridency and red lines some authors are drawing around A.I. to be extremely counterproductive.

Yes, a lot about A.I. is bad

Look, I’m not personally a fan of our increasingly authoritarian and creepy tech oligarchy. I’m deeply unnerved by where things are headed. There are tons of great arguments that nearly everything about A.I. is quite bad indeed:

• Generative A.I. is the product of colossal theft. Because of the quirks of copyright law, book authors are actually uniquely positioned to profit at least somewhat from the grand theft the tech companies perpetrated to produce these chatbots. But nearly everything ever written on the internet got hoovered up, meaning A.I. chatbots exist due to the unpaid labor of nearly everyone. We should all own a stake in these companies.
• Environmental concerns are real. When massive data centers aren’t raising power costs for local communities, they’re often being powered by their own gas-burning sources. As the effects of climate change are everywhere, there’s a risk that this hyper-scaling is being done in a way that will be even more destructive to the environment than our inadequate current path.
• A.I. is destroying the open web. Sites like this one exist because Google historically drove traffic to sites like mine so people could find the information they needed. Now people stay in the closed loop of Gemini, meaning they might get my advice regurgitated to them without knowing I even exist. Which is quite bad for my business and the future of this site.
• A.I. is bad for human creativity. There are quite a few things A.I. is good at. At least right now, creativity is not one of those things. While it can be coaxed into producing some useful utilitarian stuff, it’s not good at generating creativity, and it’s not good at editing. We risk a great dumbing down and cheapening of art by leaning on A.I. crutches to produce stories.

I’m not a tech overlord fanboy by any stretch of the imagination. I am sympathetic to everyone who wants to resist all of this.

I just don’t think the best way to resist, or to adapt to the world that’s coming, is to thrust one’s head in the sand and draw unrealistic red lines.

The A.I. era is here whether we like it or not

A big reason red lines against A.I. are untenable: the line between simply using the internet, using your phone, and using A.I. is increasingly blurry, and it’s only going to get blurrier.

“I don’t want an agent who uses A.I…” The premise of the statement is going to feel increasingly bewildering.

Where, exactly, is someone meant to draw a red line against A.I. use?

Sure, right now you can turn off Apple Intelligence and you can use non-A.I. alternatives to Google and you can eschew ChatGPT, Gemini, and Claude, but A.I. is going to increasingly be interwoven with our basic interactions with the internet and our phones. The new version of iOS coming in the fall is going to put A.I. front and center. Everywhere you turn, A.I. integrations are going to greet you (and pander to you).

Yes, this is being crammed down our throats without widespread support. Yes, the resulting gagging feeling is unpleasant.

But trying to escape A.I. at this point is like trying to live off of the electrical grid. Theoretically possible, but necessitating an entire reorientation of your lifestyle.

Maybe more importantly, I would want my agent to be deeply familiar with A.I.

The best guard against slop is being familiar with what A.I. writing sounds like

A.I. detection tools aren’t an exact science, and some studies have shown that Black, neurodiverse, and non-English writers are more likely to be unjustly flagged.

There’s unlikely to be a knight in shining armor that any agent, publisher, or competition judge can employ to draw the line against A.I.-assisted stories.

So don’t try. Just judge books on their merits.

But how are agents supposed to detect A.I. slop writing if they’re not “allowed” by authors to use A.I.?

The “tells” when it comes to A.I. writing abound, not least the pervasive “it’s not X, it’s Y” frame, which has been surprisingly difficult for the A.I. overlords to eradicate. A.I. is also fond of nonsensical metaphors, aimless stage direction, and awkwardly placed em dashes.

I want my agent to be familiarizing herself with what A.I. can and can’t do.

And if she wants to refine our list of comp titles by running my plot description through Claude? I just don’t think that search is going to be the difference between paradise and the apocalypse…

So how do you fight back?

I understand people are angry at the direction the country is headed, they feel ripped off that a handful of tech companies are raking in billions on the backs of our collective labor, and it doesn’t feel like any of us have had a say in the direction all this is going. Our current leaders are too busy insider trading and running crypto scams to do anything to help broader society. I’m certainly anxious about the shift away from the open web and what it means for this site and community.

And look, if you’re angry about all of this and feel compelled to call out practices you don’t like, who am I to stop you? I’m not trying to silence anyone’s voice.

Still, I personally think writers should consider spending less time playing the enforcement game, and more time playing the “support the world you want to see” game. (And I’m reminding myself of that today as much as I’m reminding anyone else.)

Litmus tests get lost in the slippery slope. Red lines get drowned. Internet freak-outs, even justified ones, nearly always overshoot the “crime.”

There’s more we can affirmatively do to build a better world in the face of all of this.

Don’t like how Amazon is positively drowning in A.I. slop books? Buy your books at your local bookstore, or at Bookshop.org if you don’t have one of those.

Boost your favorite human authors. Instead of plugging your novel or query into Claude, find a human editor. Support politicians fighting for equality and slowing down climate change.

Shouting at a literary agent online because she tested out Gemini for comp titles is just not the way. There are so many things we can do every day to build a better world rather than shaking our fists at tsunamis.

Need help with your book? I’m available for manuscript edits, query critiques, and coaching!

For my best advice, check out my online classes, my guide to writing a novel, and my guide to publishing a book.

And if you like this post: subscribe to my newsletter!

Art: Detail of Louis Pasteur by Albert Edelfelt