Is AI a Threat to the Addiction Recovery World?
Why the rapid proliferation of large language models scares me.
I’m a couple of years late to the AI party.
When ChatGPT first started appearing in the news every day, I tried it and was thoroughly unimpressed. Although it imitated human writing surprisingly well, it didn’t produce good writing. I could easily tell it was computer-generated, and it struck me as particularly bad at constructing cohesive long-form essays.
I wrote generative AI off as an over-hyped fad and did my best to forget about it.
However, over these past couple of years, ChatGPT and its ilk have become increasingly hard to ignore. Even when AI writing falls short of human writing, it’s often close enough to get by. I found this out the hard way when I lost one of my biggest business-writing clients to ChatGPT. Sure, the AI still wasn’t writing quite as well as I do, but it was saving the client $400 per report, so I can’t blame them for deciding the trade-off was worth it.
That happened about a year ago, and AI has only become more powerful in the meantime. Now I’ve started to wonder, will AI replace me as a sobriety writer too?
On its surface, the idea sounds absurd and even offensive. I write about my personal experiences with addiction and recovery, revealing intimate details of what it was like to struggle with alcoholism. My writing is meaningful precisely because it comes from my true, lived experiences.
And yet, by now we all know that AI could do a decent job imitating this type of writing. Large language models like ChatGPT are trained on publicly available written material, which includes more addiction and sobriety blogs than a human could read in a lifetime. These models can generate writing about addiction just as easily as they can create business writing, short stories, or anything else.
Of course, what they write won’t be real—it’s not as if the AI has ever experienced addiction—but it will likely be realistic enough to be believed as real. With perhaps just a few minutes of editing a week, someone could publish an AI-generated sobriety blog that comes across just as real as anything that I’ve ever written.
Why?
Even if AI can easily generate fake sobriety content, you might be wondering: Why would anyone bother to do so? The answer is simple—for money.
Like it or not, sobriety is a huge business.
This newsletter that you’re currently reading earns me about $50 a week. Back when I was writing more often and my blog was more popular, I often earned several hundred dollars a week. And I’m on the very low end of what people make from this kind of writing because I’ve never promoted any services or products.
The real money comes when sobriety writing is used for marketing. Blogs about sobriety can bring people to rehab programs, sober event planners, and alcohol-free drinks. When the writing is used in this way, it can become a real business.
I’m not saying that to criticize people who earn money from the sobriety world. In my book, it’s OK to earn money while doing something that helps the world.
However, it’s also important to recognize that the profit motive is out there and that for some people, it can lead to nefarious, immoral actions.
Although the vast majority of people working in rehab or planning sober events are driven by a true desire to help addicts, there’s no question that a small few are trying to take advantage of addicts and make a quick buck.
I worry that with the rise of AI, the number of people taking advantage of addicts will multiply.
It’s going to become very easy for people to generate massive amounts of content about addiction and to drown out the people who are writing about their real experiences. Newsletters like this one are likely going to be left in the dust.
The biggest downside of this is that the AI-generated content is simply not going to be as helpful to people who are trying to quit drinking. In part, because it still won’t be capable of capturing the real, lived experiences of addicts. However, even more importantly, because every AI-written piece of sobriety content will be focused first and foremost on marketing.
Again, I’m not saying that it’s inherently evil to make money in the sobriety world. However, AI-generated content is likely to be only about making money, with no real intent to help addicts whatsoever.
What’s the Solution?
As much as I hate to say this, I think that AI is going to kill off sobriety blogs and newsletters like mine. As AI continues to improve, it’s going to become indistinguishable from human writing. There are going to be entire fake sober personalities that are 100% driven by AI.
I wouldn’t be at all surprised to find out that some of the popular sobriety blogs have already turned to AI. Even before AI existed, I’ve had the suspicion that some sobriety blogs were written by people who weren’t actually sober. Whether that was the case or not, it’s headed in that direction.
I can promise that I’ll never turn to AI to write this newsletter, but nonetheless, I expect to eventually be drowned out by the swell of computer-generated content.
I think that for struggling addicts, the only real solution is to focus on in-person meetings. For decades, in-person meetings were the only real way to connect with other addicts, and I think that as AI clogs up the internet, we’re headed back in that direction.
These meetings have their downside, but they’ll remain the one reliable way to meet other alcoholics face to face—to hear about another person’s experience directly from that person.
I hope I’m wrong and that sobriety blogs and newsletters will remain a helpful resource for many years to come. However, if I’m right, I take solace in remembering that traditional, in-person, peer recovery groups will continue to exist.