AI Isn’t for Everyone—And That’s Okay

By Sean Brenner

In today’s rapidly changing professional landscape, one question that gets asked ad nauseum is “How are you integrating AI into your workflow?” Everyone has a different answer to that question, and mine is blessedly simple: 

I generally don’t.

There is an exception to that statement: I use an AI-based video transcriber for one client. My role is to create and refine transcripts of multi-hour livestreams, and I use a tool to generate a rough transcript that I can then clean up and make usable. 

But when it comes to the tools most people think of as AI—the chatbots, the image generators, the search tools, the agents—I don’t use any of them. With the exception of that one client’s projects, my process is entirely AI free, and I don’t see that changing any time soon. 

A lot of the time, when people who don’t use AI talk about why, they focus on moral or personal reasons. Although those are entirely valid justifications for not using AI, I’m not going to be talking about that here. While I definitely have my issues with AI, a large part of why I use it isn’t personal, but practical. There are many practical reasons for not using AI, and in this post, I’ll discuss some of the most important.

AI Can’t Do Everything

As most people who use AI know already, not even the most sophisticated AI tools can do everything. This is particularly impactful in fields like editing and writing. While outsiders might think AI could automate the whole field, AI’s actual capabilities often prove lacking. 

An AI tool can generate topics, draft an outline, or speed up research (so long as you research responsibly and fact-check everything it says). But when it comes to writing or editing, there are many parts of the process where a human is still necessary if you want quality. 

For example, as I’ve written about before, genuine writing isn’t something AI is capable of. While it can generate a suitable-enough imitation for something like a quick email to your boss, AI-generated text is substandard for anything beyond that level.

This is a crucial point to keep in mind when considering whether to integrate AI into your workflow: You have to be aware of what these tools can actually do. AI works best as a helper, not a replacement. Even then, sometimes it just isn’t very helpful.

AI Doesn’t Fit My Workflow

Narrowing down what an AI tool can do shouldn’t be the only thing that goes into whether you decide to use it. You also want to consider how well those capabilities pair with how you do your work.

In my case, the way I write makes it very difficult, and often downright counterproductive, to integrate AI into my workflow. A large part of my writing process happens while doing the very tasks that AI is best suited to automate: topic creation, research, and outlining. For me, shaping and refining the concept I have for the manuscript–the main part of the writing process AI can’t do–happens as a byproduct of doing those basic, mechanical tasks. 

Far from making my workflow more efficient, cutting out that part of the process hampers my ability to actually write. For me, the many steps of the writing process blend together in ways that just make it impractical to automate with AI. 

Everyone’s workflow is different, and some people simply work better without AI.

Free Models Are Ineffective at Best

Affordability can be another major concern. The most effective AI models are all paid products. The freely available AI models often don’t cut it for professional use, with low-quality output and a strong tendency to hallucinate. Also, there are often strings attached to the free versions, such as not being able to opt out of letting the company use your data for training purposes.

If you’re going to use AI for anything professional, you need a paid model, and for many people, that expense just isn’t worthwhile. Maybe you wouldn’t use it enough to justify the cost, or maybe your bottom line is just too tight at the moment. Whatever the reason, if you can’t afford a paid model, you’re usually better off not using AI at all.  

AI Sycophancy Is Dangerous

Even the most advanced AI models can be dangerously sycophantic, which has led some people to rely on them too much. Overreliance on a yes-bot can affect your decision-making for the worse, potentially damaging both the quality of your work and, for freelancers, your business. Being able to consult an AI model that will always tell you what you want to hear discourages critical thinking, and outsourcing decision-making to a machine that cannot actually think can be disastrous, as some CEOs have recently learned to their misfortune.

There’s also the increasingly widespread phenomena of AI-driven psychosis and even AI-driven suicide to consider. For people with mental health struggles or difficulty picking up on AI sycophancy, it might be in their best interests to limit or avoid certain uses of AI, especially those that involve direct interaction with a chatbot or chatbot-like interface. Even if you don’t have preexisting mental health struggles, you still run the risk of AI psychosis if you start outsourcing all your critical thinking to AI.

External Factors May Forbid It

In many fields, there are also external pressures to not use AI that you can’t ignore. For example, in the US, AI-generated text is completely ineligible for copyright. This also applies to text that’s mostly but not exclusively AI generated, so just making a few edits won’t cut it. If a manuscript you’re working on needs to be protected by copyright when it’s published, this heavily restricts your use of AI.

In other fields, there’s extensive social pressure against AI use. For example, I primarily write scripts for YouTube videos, and in many corners of YouTube, the slightest whiff of AI use will have people drawing their torches and pitchforks. They’ll leave angry comments, they’ll stop watching your content, and your reputation will be tarnished in a way that’s very hard to repair. In those sorts of fields, avoiding AI use—or at least keeping it as inconspicuous as possible—is a more important consideration than how AI might speed up your workflow.

Everyone’s conditions are different, and there are many conditions that necessarily exclude or limit AI use. Maybe it’s an external factor beyond your control, or maybe AI just doesn’t work well for you. Whatever your reasons, not using AI isn’t the impractical, career-ending choice it’s sometimes made out to be. The way you work best is the way you should be working.