To AI or Not to AI?

What a silly question...

Anyone else have a severe case of whiplash lately? Am I supposed to streamline and automate EVERYTHING, or will I be punished and thrown out for using the tools at my disposal?

While building my fractional company and this newsletter, I’ve been clear with my audience: I LOVE AI. I feel fortunate that it’s at my fingertips while I launched this portfolio career of mine. Truly couldn’t have seen the success I’ve seen in the last year without the help of these tools.

Last week I sat down drafted, edited, and posted my newsletter. Breathed a sigh of relief. Then went to my Substack notes feed and immediately saw a post from Wes Pearce saying Substack rolled out AI detection.

WTF??? My heart dropped. Panic ensued. I hadn’t received any heads-up this was happening. What does this mean? Where does that show up? Who sees that? Can I turn it off? Will I be punished for how I just used AI?

I felt like I was about to be outed as a fraud to my audience. My first thought was far from “Oh wow this is an interesting and helpful feature!” It was “What if my readers assume everything I do is GPT with my name at the top?” That is incredibly uncomfortable when you’ve spent countless hours of energy and heart working to build trust. This complete flip of a belief that you’re not valuable if you use AI was foreign and felt like an attack.

As the dialogue progressed and my nervous system chilled out, I realized I would love to be super open about how I use AI in my work. So I want to take this opportunity to offer a deeper peek behind the curtain into how I use the tools. If you’ve been here a while, you know my brand is transparency and human-first. Even in my LinkedIn banner, I make it clear that I am about human-first, AI-powered systems. I consider this to be a huge benefit to my clients and readers. AI is NOT going anywhere. I strongly believe if you don’t adapt and use it in your practices, you will be left behind. I still believe this after feeling like I was stripped down naked and the Substack lords yelled “Shame” upon me. (God, does being a Game of Thrones fan just age me now?)

Anyway, I am going to share my process for using AI in my content here. The process is a lot like working with another human in the room, but as a solo, I don’t have that luxury. When I sit down to write my newsletter for the week, there are a couple of ways it goes.

Sometimes I have the fully fleshed-out idea for an article.

Sometimes I have a vague idea of where I want to go with an article

Sometimes I sit down with severe writer’s block and stare at the screen for a hot minute.

Whatever the case, I usually will still have my GPT or Claude (depending on who is on my good side that day) interview me. Ask me about what I am seeing with clients, what’s being talked about in my world, and what I am feeling lately. For this piece of the process, I like using Wispr to speak it out. It feels more natural.

Once I get that out, I outline, draft, rewrite, edit. Then I hand it back to my robot buddy for feedback.

The journalism kid in me despises being called out for grammar issues, so I make sure that it’s checking me on my grammar. (I’m not perfect but damn my brain sure thinks I should be.) If there’s research that I need to do for the piece, I work back and forth with Claude to help me surface that. I still go through and click on the sources and articles that Claude’s giving me to verify because it’s not always accurate. It’s like your well-meaning and hard-working, but still rather young, intern.

Then, at the end of all this, I’ll put it on Substack and do one last pass on it myself. Okay, let’s be real: 10 passes because aren’t we all afraid to commit to clicking the send button only to send something horrifically wrong? (Anyone else previewing all 3 formats a stupid amount of times before send? Anyone??) Eventually, I finally fire away.

The stories, frameworks, and witty quips 😬 start in my brain first. But, YES, the AI does help me clean this up and give the final product you’re reading. I don’t know what grade I’m going to get from that detector, but it’s fair to say it’s not 100% human these days. It’s also not close to 100% AI.

Which is why I am still not sure what the detector is measuring. I’ve read thoughtful arguments from both the supporters and the haters. Maybe this removes some of the low-effort “AI slop” out there. And I am all for that! I don’t think there is a single person out there asking for more generic content. But I keep finding myself coming back to the negatives. False negatives. False positives. Then something I didn’t even think about before: there is a whole set of people out there who speak English as a second language, and they’ve suddenly been able to access a whole new market. This feels like a punishment to them. My friend, Dr Sam Illingworth speaks about this extensively in his publication, Slow AI.

I'm just not convinced we've figured out how to measure "human" in a way that's actually fair.

Other AI tools I am using to streamline my business because it doesn’t stop here:

Claude

ChatGPT

Whispr

Granola

Grammarly

Gamma

Lovable

Probably five more next month because I have absolutely no self-control when a new AI tool drops.

If AI really isn’t your thing?

That’s cool. You can still hang with me.