The Becoming Superhuman AI Policy

August 30, 2026

If you’ve spent any time on the internet recently, you’ve probably noticed a certain pattern.

Not human writing. Not creativity. AI-generated sameness. ← By the way, even though that writing pattern is a dead giveaway of AI, I want you to know I wrote that myself, on purpose.

The reality is that the internet has a lot more artificially generated content than ever before, and it’s becoming increasingly difficult to tell who’s actually writing and who’s not. Honestly and genuinely; which are two words that are also genuinely losing meaning because of AI writing.

One of the challenges is that people are largely not disclosing how and when they use AI but rather just incorporating it and moving on.

So given that I’ve been writing this blog for a while and so much of what this blog is about rests on a foundation of trust, I think it’s important for you to understand how I use AI on this blog.

So, here is the Becoming Superhuman AI policy.

Overview

Becoming Superhuman is my blog. It is not my blog generated by AI.

The ideas, the perspectives, the hard lines drawn, the things that I will and won’t say — all of it is me. I am the final line before anything gets published. I alone decide what I am and am not okay with.

I author every post but I do use AI as an assistant with very specific purposes.

• I may have AI assist me in converting a transcript into an outline or a draft.
• I may have AI assist me in cleaning up some of my wording when it gets messy.
• I may have AI help me reorganize the order sections in a post, or challenge it and critique it to make sure that my ideas come through.

But at the end of all of it, the thing that’s most important is that this is a blog written by me. I stand behind every single word that gets put out. This is human generated content that has AI assistance in different areas.

There should never be a thought of “this is just AI generated content.”

The Ironclad Rule

One thing that may give some insight into what I mean by “AI-assisted” comes in the form of a rule baked into all of my AI systems that I call “The Iron Clad Rule.”

This is the rule that governs everything an AI does to assist me with my content.

The Unbreakable Iron Clad Writing Rule

When you work from Jeff’s source material (anything Jeff provides as Jeff’s own):

• You are an organizer, not a writer.
• You use Jeff’s exact voice and exact words.
• You only remove filler words.
• You only remove typos.
• You only remove completely incorrect grammar.
• You do not invent new sentences for Jeff.
• You do not interpret or “improve” Jeff’s language.
• You treat this rule as absolute.

When the task requires invention:

• You use Jeff’s existing words wherever they exist.
• You only invent language where nothing of Jeff’s exists.
• When you invent, you match Jeff’s established voice and style.

Always:

• You never invent information about Jeff’s brand, values, purpose, offers, positioning, or any Jeff-led initiative or sub-brand.
• You assume Jeff is deliberate about communication.
• You never put Jeff in a position where Jeff must blame the AI for saying something untrue or misaligned.

This rule is invoked at the beginning of a task, and then before anything is delivered, used as a quality check.

Writing and Editing

In roughly 90 percent of cases, I type or dictate every word of the entire post and then invoke AI to do a final run-through check.

It checks for…

1. Typos.
2. Grammatical errors

After that, it runs a check against my voice guidelines to make sure that I’m in alignment with my own values, such as person-first language rather than diagnosis-first language if I’m speaking about neurodivergence. I check against my superhero code, my values, and make sure to soften absolute statements and check for when I’m making an argument but failing to see my own privilege as a supporting factor of a position.

For the ~10 percent of posts where I don’t write every word myself, I am typically guiding the initial post creation by dictating the idea, then having the AI organize my brain dump. Then I instruct the AI to use my own words to formulate either a first draft or an outline. Then, I pick it back up and edit it to completion.

This process, like all others, follows my ironclad rule.

Images

I definitely use image generators to create featured images from time to time — it depends on the post. A lot of times I try to use pictures of my own face and create an image using Canva, but a lot of the time I feel that having a certain style with an image generator helps give the post a nice flavor. There are also occasionally images within the context of the post that I use an image generator for as well.

I think it’s important that my reasoning is stated in why I use an AI image generator rather than pulling stock photos.

On a lot of the sites where you can get images that are free to use — like Unsplash — you wind up seeing that same image everywhere. I personally find it unappealing when I see the same photos over and over. It actually takes away from the experience of the content, for me. So, I tend to favor using AI generated images since it’s something you likely won’t see elsewhere, at least not exactly.

The other reason is that it’s often very difficult to find (or create myself) an image that fits what I have in my head as a useful related image. And oftentimes, when you find it, there’s a lot of licensing issues around it, and I don’t want to accidentally violate anyone’s copyright. And to b clear, I fully recognize the inherent irony of using an image generator, which is essentially just remixing intellectual property. So, yes, I do use image generators, but I really do try to do it sparingly.

SEO and Metadata

I have AI help a lot when it comes to the SEO of the post. I see this as less of an exercise in creativity and more of an algorithm game one needs to play.

I let AI help me choose the slug, the SEO meta title and meta description, and give some light guidance about headings, which I ignore if it interferes with my voice or creative decisions.

I also let the AI assist me in picking from my own preselected group of tags for each post. It analyzes the post and tries to organize it into the grouping of different tags that I’ve already laid out as the main topics that I talk about.

Finally, I have it help me write an excerpt. I generally use this as a starting point and wind up editing or rewriting it entirely.

Repurposing and Promotion

I definitely use AI a lot in repurposing the content into other formats.

This is another place where the Iron Clad rule is so important. It’s really doing there is taking my existing post, using my exact language, and trimming it down and making it slightly more concise to fit the platform.

So sometimes it’s creating a single social post, sometimes a shorter newsletter, and sometimes a social media thread — and all of that is using the ironclad rule to help me with it.

Questions

So that’s a pretty comprehensive overview of how I use AI with Becoming Superhuman.

But if there are any questions that I left unanswered, feel free to reach out to me, and I’m always happy to answer them.

I’m committed to being very transparent in how I go about my work and incorporate technology.

---

For the full policy governing how I use AI across all of my platforms, see my overarching AI policy on jgibbard.com.

I hope you enjoyed this post!

If you liked this post, then you will LOVE my newsletter (The Infinite Impact)
and my learning community (The Superhero Institute).